suzanne strasser · 2019-11-07 · make business decisions. oracle data integrator 11g and oracle...

The Magazine for the Rocky Mountain Oracle Users Group • Vol 72 • Fall 2013

Rocky Mountain Oracle Users GroupPO Box 621942Littleton, CO 80162

Change Service Requested

Non-ProfitOrganization

U.S. Postage PaidSan Dimas, CAPermit No. 410

Member Focus - Lena CohenBoard Focus - Bobby Curtis

OBIEE Tips & Techniques - Dan Hotka

Suzanne StrasserOEM 12c Metrics Data

Marcelle KratochvilMyths And Multimedia in Oracle

Michael RaineyGolden Gate & Oracle Data Integrator

Kent GrazianoMaking Data Modeling Fun

2 SQL>UPDATE • Fall 2013

ExplorE In-DEmanD DEgrEEs | rEquEst aDDItIonal InformatIon | apply now

> Classes start soon

Be InFlUentIal.

on yoursCHEDulE

Busy adults choose regis university for the academic rigor and flexibility.

> regionally accredited and nationally ranked> accelerated 5- and 8-week classes> online and campus-based learning> multiple start dates throughout the year

CHangE yourworlD

sharon combines her technology background with her regis degree to help empower her own community. How will you be influential?

a sCHool ofInfluEnCE

the school of Computer & Information sciences in regis University’s College for Professional studies offers four undergraduate and six graduate degrees that immerse students in the latest technologies and software. students benefit from our state-of-the-art learning environments, experienced faculty and strong industry relationships.

CPs.regis.edu/beinfluential | 800.659.6810 | sharon Joan Whitehair m.s. in Database technologies

Learn more at:CPs.regis.edu/beinfluential

It’s all about how to allow

themselves...

people toBEttEr

Volunteer for Training Days 2014 !Room Ambassadors Receive Full Registration

Registration Ambassadors Receive 1/2 Registration FeeFor questions about Training Days, contact [email protected].

February 5-7, 2014Information & Updates Available On http://www.rmoug.org

Contact [email protected] to Volunteer

SQL>UPDATE • Fall 2013 3

On the Cover:“Fall colors year round.” The cover photo of Red Mountains was taken in July of 2013 from a rocky

road between Ouray and Silverton, CO, near the area called Corkscrew Gulch. The entire area is a popular destination for 4-wheel driving. The photographer was Stefan Zielinski, an Oracle DBA (since 2005) and a long time C and PL/SQL developer. He currently works for Dell Services.

c o n t e n t s

Editor & dirEctorPat Van Buskirk

[email protected]

SUBScriPtioNS & tEchNical adviSor

Heidi [email protected]

coNtriBUtiNg WritErSLena CohenBobby CurtisKent Graziano

Dan HotkaMarcelle Kratochvil

Michael RaineySuzanne Strasser

SQL>Update is published quarterly by

Rocky Mountain Oracle Users Group4869 S Coors Ct

Morrison, CO 80465 (303) 929-6266

Email: [email protected]

Please submit advertising and editorial mate-rial to the [email protected]. RMOUG reserves the right to edit, revise or reject all mate-rial submitted for publication. Liability for errors in ads will not be greater than the cost of the advertising space. ©2013 Rocky Mountain Oracle Users Group

features

monthly features

4 Letter From Our President5 Stan Yellott Scholarship Fund28 RMOUG Board of Directors30 RMOUG Calendar of Events31 November 2013 QEW

A special Thank You to Heidi Kuhn, Tim Gorman and Kellyn Pot’Vin, without whose continuing help, this publication would not be possible.

6 Golden Gate & Oracle Integrator by Michael Rainey

A Perfect Match

10 Exposing the Myths by Marcelle Kratochvil Managing Multimedia in Oracle

22 Five Ways to Make Data Modeling Fun by Kent Graziano

Try These Games!

14 Querying Metrics Data by Suzanne Strasser OEM 12c Management Repository Views

departments

19 OBIEE Tips & Techniques Dan Hotka24 RMOUG Member Focus Lena Cohen26 RMOUG Board Focus Bobby Curtis


Last month on Friday August 9th, RMOUG had its Summer QEW (quarterly educational workshop) at Elitch Gardens

Theme & Water Park in downtown Denver. We had a full agenda of presentations from Oracle Corporation in the form of their “Database12c Technology Day”, part of a road-show of techdays all around the country.

From The President Tim Gorman

The agenda was impressive, covering all of the new features of Database12c, including the following sessions (along with the links to the official Oracle presentation downloads)...

Database12c Security (http://www.oracle.com/us/dm/secu-•rity-inside-out-12c-1983117.pdf)Database12c Availability (http://www.oracle.com/us/dm/•maximize-availability-with-12c-1983119.pdf)Database12c Automatic Data Optimization (http://•www.oracle.com/us/dm/automatic-information-lifecycle-12c-1983121.pdf)Database12c Consolidation (http://www.oracle.com/us/dm/•simplify-consolidation-with-12c-1983122.pdf)Database12c and Exadata (http://www.oracle.com/us/dm/•drive-dbaas-with-exadata-x3-pdf-1983126.pdf)Managing Database12c with Enterprise Manager 12c •(http://www.oracle.com/us/dm/oracle-enterprise-manager-12c-1983127.pdf)

The program ran from 8:30am until 2:00pm and attracted 150 attendees and about 60 friends and family. Each attendee received free parking at Elitch’s, lunch was provided, and a free all-day park pass, while friends and family were able to buy an all-day


park pass for $15/per-person, courtesy of Oracle Corporation, Intel Corporation, and RMOUG.

Presentations can be downloaded either from the official Oracle links shown above, or from the 2013 QEW webpage on the RMOUG website at “http://www.rmoug.org/workshops/qew-2013-presentation-summaries-and-papers/”. The presentations on the RMOUG website are those actually presented at the QEW at Elitch’s, since the RMOUG presenters made modifications to the official Oracle presentations.

The word is that another big Oracle TechDay road-show is starting later this year with a stop in Denver in early December 2013.

This will follow closely after the Autumn QEW on Friday November 15th at the Oracle campus in the Denver Technology Center, and just two months prior to the big annual “Training Days 2014” conference in downtown Denver at the Colorado Convention Center. There’s a lot going on at RMOUG, so keep an eye on our website (www.RMOUG.org) for the latest in events and learning opportunities.

RMOUG Scholarship Mission

To provide educational opportunities to members of the organization about the information technology indus-try in general, and in particular the technology of Oracle Corporation to include databases, storage, networking and application development, specifically the products and services of the Oracle Corporation.

To collect, post and distribute information about Oracle technologies and other related technologies to members.

To provide members with the ability to help their peers maximize their knowledge and skills working with products in information technology and in particular Oracle products.

To provide a consolidated channel of communication, con-veying needs, concerns, and suggestions, for members of the organization, to Oracle Corporation and other vendor corpora-tions involved with Oracle related technology.

To encourage members to present their information technology experiences using Oracle and other products and services.

To provide a consolidated channel of communication between members of the RMOUG and other communities in related information technology industries.

To promote educational opportunities for students of information technology through directed funding and services for educational purposes.

RMOUG is committed to supporting others in the pursuit of technical knowledge.

The Scholarship Fund started in 2001 to encourage future IT professional in their efforts to broaden their knowledge. In 2007, RMOUG voted to rename the scholarship fund to honor the memory of Stan Yellott. Stan was a long time member of RMOUG where he supported the user community by serving on the RMOUG board. Stan focused on expanding Oracle educational opportunities. Stan’s vision was to include high school and college students as the next generation of IT pro-fessionals.

For Details, Visit the RMOUG Websitewww.rmoug.org

Stan Yellott Scholarship Fund


Golden Gate & Oracle Data IntegratorA Perfect Match

by Michael Rainey, Principal Consultant, Rittman Mead

Many organizations are planning a real-time data warehouse implementation, providing their end users the ability to rapidly make business decisions. Oracle Data Integrator 11g and Oracle GoldenGate 11g both excel as individual products, but paired together they form a perfect match for real-time data warehous-ing. Following Oracle’s Reference Architecture for Information Management and Big Data, this article will provide best practices on how to configure, implement, and process data in real-time using Oracle Data Integrator and GoldenGate.

Oracle’s Reference Architecture for Information Management

The Reference Architecture for Information Management and Big Data is the latest version of Oracle’s approach to data management. There are three layers to the architecture: Staging, Foundation, and Access and Performance. This architecture will be used throughout the article as the basis for the real-time data warehouse.

The Staging layer acts as the “landing pad” for incoming source data, with an exact copy of the current source table and / or change data. The Foundation layer also retains a copy of the source tables, but rather than storing only the current rows from each source, the entire transactional history is stored. The history is tracked with audit columns such as System Change Number (SCN), commit date, and change type (insert, update, or delete). Finally, the Access and Performance layer is where the traditional star schemas are created and business intelligence tools, such as Oracle Business Intelligence Enterprise Edition (OBIEE), will con-nect and access the data warehouse data.

The key to the architecture is the aptly named Foundation layer. This layer is setup to be isolated from source system require-

ments, allowing the historical data to be stored as the source changes over time. The historical data can be used to reload the target star schema should new attributes or measures be added. It can also be used to support reporting against transactional schemas. For example, creating a logical OBIEE model against the Foundation layer can provide the end user a prototype of a report, further defining the requirements before the final ETL is constructed.

Oracle GoldenGate and Oracle Data IntegratorOracle GoldenGate is the standard Oracle product for data

replication, providing log-based change data capture, distribution, and delivery in real-time.

GoldenGate captures transactional data changes from the source database redo log and loads the changes into its own log file, called a Trail File, using a platform-independent universal data format. The Extract process understands the schemas and tables from which to capture changes based on the configuration set in the Extract parameter file. The data is then read from the Source Trail File and moved across the network to the Target Trail File using a process called a Data Pump, also driven by a parameter file. Finally, the transactions are loaded into the target database tables using the Replicat parameter file configuration, which maps source tables and columns to their target. The entire process occurs with sub-second latency and minimal impact to the source and target systems.

Oracle Data Integrator (ODI) is Oracle’s strategic product for data integration. Using the E-LT (Extract, Load, and Transform) methodology to gather data from the source, load it into a staging area, and then perform transformations while loading the final tar-get, ODI uses the power of the database to perform the work rather than an additional ETL engine. Mappings are generated as ODI Interfaces and use customizable code templates, called Knowledge Modules (KMs), to generate the loading and integration code during execution. The power behind ODI Interfaces is the use of KMs that generate native, set-based database code.


Oracle Data Integrator also has its own form of change data capture (CDC), which is implemented using a Journalized Knowledge Module (JKM). The JKM generates the CDC infra-structure, creating journal tables that store the change rows and views that provide access to the change data for consumption by Interfaces. The ODI CDC capture process can be implemented using triggers on tables, last change timestamps on rows, or mining of the database logs via Oracle Streams or Oracle GoldenGate.

The integration between GoldenGate and ODI begins with the Journalized Knowledge Module, “JKM Oracle to Oracle Consistent (OGG)”. The JKM, delivered with ODI, will create the GoldenGate parameter files, provide instructions on how to imple-ment GoldenGate, and generate the ODI CDC tables and views. The JKM is added to the Model and uses ODI Datastores, a logical representation of the database tables, to generate the GoldenGate parameter file mappings. This alone can save quite a bit of manual work and can reduce typos caused by human error.

Loading the Reference Data Warehouse in Real-Time

Oracle GoldenGate and Oracle Data Integrator will be used to load each layer of the Oracle Reference Architecture for Information Management. GoldenGate is implemented to capture changes on the source and populate the Staging layer database tables. GoldenGate will also load the Foundation layer directly from the source, as it is built to capture changes once, and then distribute the change rows to multiple locations in parallel without any additional impact to the source system. The need for additional ETL mappings to incre-mentally load Foundation from Staging is eliminated, along with the added maintenance. Finally, ODI Interfaces are built to execute ETL mappings that will load the Access and Performance layer.

The Journalizing Knowledge Module “JKM Oracle to Oracle Consistent (OGG)”, which will be applied to the Staging and

Foundation Models, must be slightly modified to allow the capture of historical transactional data in the Foundation schema. The addition of the parameter “INSERTALLRECORDS” to the Replicat parameter file loading the Foundation schema will apply every transaction as an insert into the target, regardless of whether it was an insert, update, or delete on the source. The addition of data warehouse audit columns, described below, will allow the tracking of transactional history in each table.

EDW_SCN (System Change Number) EDW_COMMIT_TIMESTAMP (when the change was committed) EDW_TRANS_TYPE (DML type for the transaction: insert, update, delete)

Journalizing within Oracle Data Integrator allows ETL devel-opers to use the change records, stored in the J$ change tables, as the source of the ODI Interface. The JKM will setup GoldenGate processes to replicate the change data from the source to the change tables. When a change occurs in the source, the primary key and Window ID, which essentially represents the source SCN, are stored in the change table. Using the change data allows for near real-time processing of ETL mappings when loading downstream data warehouse tables.

To setup the integration between ODI and GoldenGate, begin by creating the ODI Models and Datastores for the Staging and Foundation tables. A standard practice is to reverse engineer the source tables first, and then copy them to the Staging and Foundation Models. This approach will ensure the column names and data types remain consistent with the source. Next, create the additional data warehouse audit columns in each of the Foundation Datastores manually one-by-one or in a batch using a Groovy script and the ODI SDK.

Once the Models are set, apply the customized “JKM Oracle to Oracle Consistent (OGG)” to the Staging and Foundation Models and set the options appropriately. These options provide the JKM information about the GoldenGate installation on both the source and target servers, as well as naming conventions and other meta-data. Several of the options are highlighted below.

LOCAL_TEMP_DIR: Local path for generated parameter filesSRC_LSCHEMA: Source Logical SchemaSRC_DB_USER: Source GoldenGate userSRC_DB_PASSWORD: Source GoldenGate passwordSRC_OGG_PATH: Source GoldenGate install pathSRC_SETUP_OGG_PROCESSES: Setup extract files if trueSTG_HOSTNAME: Target server hostnameSTG_MANAGER_PORT: Target GoldenGate install portSTG_OGG_OBJECT_GROUP: Replicat file name STG_OGG_PATH: Target GoldenGate install pathENABLE_ODI_CDC: Setup the ODI CDC framework if true

From the tree structure under the Model, right-click each Datastore that will be replicated and choose “Add to CDC”. Then, choose “Start Journal” at the Model level. Based on the options set in the JKM, the Start Journal process will create the ODI CDC framework (“J$” change tables, “JV$” change views, etc.) and gen-erate the GoldenGate parameter files, configuration batch scripts (called Obey files), and instructions on how to complete the setup.


Browse to the temp directory indicated in the JKM options, open the Readme.txt file, and follow the instructions to complete the GoldenGate implementation. One important thing to note is that the Obey scripts generated by the JKM contain start commands for the extract, pump, and Replicat processes. It is recommended that these lines are removed or commented out, as an initial load of source data to the target will need to be performed prior to starting replication.

To perform the initial load, first start the GoldenGate extract and Pump processes on the source. This will begin the extract and replication of source data changes over to the target trail file. Next, choose a batch load method, such as Oracle Data Pump, and move the source data to the target using the flashback query approach, as of a specific SCN. While GoldenGate does provide an initial load capability that can also be generated via the ODI “JKM Oracle to Oracle Consistent (OGG)”, it is very slow at moving the data from source to target and therefore is not recommended. Once the data has been loaded, start the Replicat on the target after the SCN cap-tured during the initial load process, ensuring no transactions will be duplicated or skipped.

GGSCI> start replicat ODIT1A aftercsn 123456

GoldenGate is now setup and replicating source changes to the Staging and Foundation schema tables, and the “J$” change tables for use with ODI change data capture.

Real-Time ETL Using Oracle Data Integrator CDCETL mappings, built as Interfaces in Oracle Data Integrator

11g, are developed to populate the Access and Performance layer. As described earlier, the Interfaces will use the ODI Change Data Capture (CDC) framework to flow data captured in the change tables through to the target. The CDC framework includes the fol-lowing:

Journals• - tables (prefixed with J$) that hold references to the change records Journalizing views• - (prefixed with JV$, JV$D) provide access to the change data by joining the journal table to the fully replicated table, and are used by IKM’s and LKM’s to access the change rowsCapture processes• - captures changed data from the source tables (Ex. Oracle GoldenGate)Subscribers• - entities that consume the changed data as a consistent set

When the change data is consumed for a specific set of tables, the “window” is extended for that change set. This will essentially freeze the set of change data by only allowing transactions between the minimum and maximum WINDOW_ID to be consumed, ensur-ing that no transactions are missed and allowing for consistency across the dataset. Once the data has been processed, the set of change data for that given window is purged from the change tables. More detailed information on ODI Change Data Capture can be found in the ODI documentation.

Using Journalized DataOnce journalizing is setup and changes are being captured,

building an Interface for real-time data warehousing can be quite simple. Create a new Interface, add the journalized Datastore as a source, and check the “Journalized data only” checkbox. This will change the source of the Interface in the generated code to be the JV$ change view rather than the actual table. The JV$ view will contain only the change rows available after the “extend win-dow” process is called. The Interface is executed and the journal is purged, setting up the process for the next run.

This process looks to be quite simple, but that is not always

the case. There is a restriction that only one journalized Datastore can be used per Interface. So in the case that both the EMPLOYEE and DEPARTMENT tables are journalized, only one is allowed to have the “Journalized data only” checkbox selected when used as a source in the Interface. Fortunately, there are several ways to work around this limitation.


STAGE_IND• – Indicates the rows that are ready to be con-sumed for the first time. In other words, these rows have not yet been “seen” by the process.CURRENT_IND• – The most recent version of that row by natural key, which, timing aside, matches the record in the source.

Using a combination of these indicators, the developer can choose to return only change rows (STAGE_IND = ‘Y’) or return all current rows (CURRENT_IND = ‘Y’). This works well for incremen-tal processing of data, but we can take it one step further. What if there is a need to reload the fact and dimension tables using the historical data in the Foundation layer schema? A view can be cre-ated which combines all three sets of data: Change rows (J$ table), Current replicated rows (Staging table), and Full transactional his-tory (Foundation table).

In this case, a bit of SQL analytics must be performed using the SCN and natural key to determine the CURRENT_IND from

Parent-Child RelationshipOften a join between two source tables is due to a parent-child

relationship: Department->Employee, Order->Order Line, etc. In this situation, and with any join really, the change data must be captured from each individual table and a join will only lookup addi-tional data from the other table when necessary to complete the full record. Unfortunately, this setup is not possible in a single Interface as was just described. So what is the workaround?

Two Interfaces must be created; both with the exact same logic, source Datastores, and target Datastore. The only difference will be which source Datastore has the “Journalized data only” option enabled. This approach will ensure that no transaction is left behind and all changes will flow through to the target with the appropriate lookup data filled in for each row.

But this might not be the best approach if there are 3 or more Datastores to join. Imagine if there were 12 source Datastores and then some logic in the where clause changes, forcing an update to the ETL code. That makes 12 Interfaces to modify just for a single change!

ODI Cookbook ExampleThere is another workaround, outlined in the book “Oracle

Data Integrator 11g Cookbook”, that takes a similar approach but with potentially less maintenance headaches. This method separates the final Interface, and all of its logic, from the activity of ensuring that each table has its changes processed along with the additional lookup data from other supporting tables. The approach looks promising (and the book is well worth the purchase), but having not yet attempted it in practice, we will not go into further detail here.

Subscription ViewsA final option for moving data from the change tables through

to the target facts and dimensions as quickly as possible can be achieved using what we call “subscription views”. The goal here is to always return a consistent set of data and to let the ETL developer make the choice as to how this is accomplished. In this solution, the “Journalized data only” checkbox is never checked for any of the source Datastores in the Interface. Instead, a view is created for each of the Staging schema tables, with a join to the ODI J$ change table. (See output in next column.)

The ETL developer can then reverse engineer the views into an ODI Model and use them as the source Datastores in the transfor-mation Interfaces. There are two flags that are added to the view to allow the developer a choice of data source. ...continued on page 13


Exposing the Myths Behind Managing Multimedia in Oracle

by Marcelle Kratochvil, Oracle ACE Director, CTO Piction

I recently read an article which made the claim that it was a well accepted fact that it was not efficient to store binary

data larger than 1mb in an Oracle database. After working for 15 years in the field of unstructured data and Oracle, I realised that if that is the view still now in the marketplace, then that view is not only completely wrong and outdated, it is also misleading and very dangerous. It’s a perception that could have been argued when using Oracle7 in the early 1990’s but from Oracle9 onwards and with major improvements into Oracle 11, this is a view that is just not correct.

I have been campaigning and pushing the idea that Oracle since 8i is much more than just a relational database. It is an object/relational database. It’s a concept many people have com-pletely forgotten. Most of the object-oriented features in Oracle seem to be dismissed or ignored by developers, including those dealing with large objects (LOBs). The conservative nature of many database administrators is exasperating, as they prefer to stay within the confines of their own comfortable relational knowledge base and steadfastly refuse to accept the performance benefits that objects can offer.

When one looks at how well Oracle itself uses the object fea-tures inside its own database, you realise these capabilities aren’t going to disappear. Having to use pure relational for development is the real myth now. Experience has shown that relational on its own just doesn’t scale and work for a lot of applications. Relational might be a solution; it’s just not the solution. With the recent push towards big data, one begins to see that the marketplace is mov-ing down the path that simple is better. Most of what big data is capable of can be easily simulated in an Oracle database (using index hash cluster storage or PL/SQL Associative Arrays). The only advantage big data has is that it scales and is designed well for a niche. Big data has exposed the fact that relational has scalability issues when dealing with certain types of data such as audit trails and non transactional data. “Big data solutions are typically used when there is lots of low level data that may not have yet been processed into transaction data”i

With unstructured data we begin to deal with multimedia data. This includes photos, video, audio and documents. There are more types in the marketplace, but these are the common four. Photos include digital images which can vary in size from a couple of kilobytes for a thumbnail image, to gigabytes for high quality images (such as georaster).

With the widespread acceptance of smart phones in the mar-ketplace, the rules for dealing with data have changed. Relational data is important but it’s not the only data. Estimates vary on the total amount of unstructured data in the marketplace, with figures

ranging from 70% to as high as 90%. “More recently, multiple ana-lysts have estimated that data will grow 800% over the next five years. Unstructured information accounts for more than 70%–80% of all data in organizations and is growing 10–50x more than struc-tured data.”ii

You would not think this is the case based on all the attention and focus given to relational. That’s important to stress. The vast majority of data in your organisation is ignored or not correctly managed.

The lesson to be learnt from the Smartphone and especially the iPad (and equivalent devices) is that the keyboard does not reign supreme anymore. Most functions on the iPad use other interac-tions to enter in data (touch, swipe, visual). With the introduction of Siri, voice recognition is there. With Skype we communicate using audio and video, and what makes these devices so popular is that they don’t have a keyboard (note how the Blackberry has declined in popularity even though it does have one).

We are in the middle of a massive change which most of us are in ignorant bliss to. How we are dealing with computers is chang-ing (for example, the release of Google glasses will usher in a new type of computer/human interaction). How we store data and man-age it is changing. We now have the need to manage audio, images and video. This isn’t happy snaps taken on our last holiday; this is information used in the day-to-day running of a business, informa-tion which is crucial. With more of these devices being used by more and more people, multimedia and the need to manage it as an object will grow, not decline.

The lesson to be learnt is that we can’t ignore this. Administrators and developer must get up to speed on this technol-ogy, understand, learn how to tame it and work efficiently with it.

The biggest lesson that needs to be learnt is that multimedia is core to how we work and deal with the information technology environment. Relational supports it. It’s crucial, it’s important, but relational data isn’t just it. It’s not the only thing. It’s needed but the bigger picture which encompasses all the data says there is more. Until we as an Oracle community begin to wake up and realise this, we are missing out on new business opportunities and most of us will be relegated to being treated as dinosaurs. Because those who do embrace and work with this technology will change the rules and then out-compete the rest of the market. The switch-over will be as dramatic and as fast as the switch from old style phones with keyboards to the newer, modern models with friendlier and simpler interaction.

The following discusses some of the myths and objections com-monly raised when working with multimedia:


Myth 1: The largest image size the Oracle database can sup-port is 1Mb.

Response: 1Mb is trivial to manage in Oracle11 using Securefiles. I have dealt with databases on Windows and Linux (even on Oracle XE) and loaded in digital photos over 2Gb in size. For video, I have loaded in digital objects over 20Gb in size. The largest digital object I have loaded in was 60Gb (which was a virtu-alization). I have even loaded an Oracle database into itself. At no point was the performance of Oracle an issue.

Myth 2: It’s easier to manage multimedia in the file system.

Response: There are two main issues with file systems. The first is that it’s hard to secure and protect them. Delete, rename or move an image on the file system and it can become orphaned if that change is not reflected back in the database. So if you use a BFILE or physical reference, then the biggest issue is managing the references and ensuring the file system is protected from everyone. Experience has shown that images will eventually be lost using this method. Whereas, if its stored in the database such an issue doesn’t exist. Using WebDAV, it’s possible to simulate a disk front end when the data is stored in the database.

The second issue is that file systems enforce a hierarchical mentality for image management. So ingrained in our thinking are hierarchical file systems that we can’t think outside the box on this and realise how constraining such a structure is. File systems don’t scale and struggle to handle more than a thousands objects in them. (Try doing directory listings in Windows or Unix when there are more than 1000 files in it. Try searching or navigating around them. You will find it’s incredibly slow, inefficient and hard to work with. I am surprised more people aren’t screaming out about how hard it is to work in this environment). A hierarchy is just one way to categorise files. There are other methods available and when you store the data in a database you can start to think true multi-dimensional.

Myth 3: Multimedia should be in the middle tier not in the database (third) tier.

Response: Experience shows that the most inefficient place to store multimedia is in the middle tier. Developers like 3 tier archi-tectures as the view is that it scales. Well, it might scale for legacy systems, but when dealing with digital objects, removing the middle tier is the most efficient method. If you put the multimedia in the middle tier, it means it’s likely not secured (digital images can then be stolen), unlikely to be properly backed up and managing and referencing them from there is difficult to do, as information about their location and status has to be sent back to the database. In the end, it becomes apparent that it’s easier to manage them when they are stored in the database and that means having the database in the middle tier. In that case, why have a middle tier?

Myth 4: Databases are designed for relational.

Response. As of Oracle8, Oracle became object/relational - a classification most people have forgotten. Relational is mathemati-cally proven and sound for dealing with structured data. The prob-lem is that 90% of data don’t fit into this category. So one needs to realise that relational isn’t the solution to all of life’s computer

problems. It forms a key component, but when you start to work with unstructured data, you soon realise the rules have changed and the relational concepts do not work. A good example is that the concept of foreign keys do not make sense with unstructured data. Even the concept of an exact match needs to be rethought. For example, when comparing two digital images how do we know they are identical? Is a byte-by-byte comparison sufficient? But what happens when you change one byte. Is it the same then? Visually they might appear identical but an exact byte-by-byte match will indicate they are not the same. Images can have watermarks hid-den within them (as pioneered by Digimarkiii) and visually appear the same. Again, a direct byte-by-byte comparison will indicate they are not the same images. Metadata can also be embedded in a digital image (using formats like EXIF or XMP). In that case, if the metadata is changed, is the digital image still the same? The answers to these questions embark on a whole new field of fuzzy logic, neural network and image matching concepts that is still in its infancy.

Experience has shown that the Oracle database can indeed manage and manipulate this data but the ability to comprehen-sively query it is still in its early stages.

Myth 5: You can’t recover when images are stored in the database.

Response: Yes you can. Oracle’s transactions work with all data, not just relational. When you store unstructured data in Oracle and do a recovery, all data is recovered equally. All data is treated equally.

Myth 6: It’s too hard to work with.

Response: The catch cry of the developer or database adminis-trator, who is stuck in their comfort zone and does not want to learn new technology. Multimedia requires learning new concepts and abandoning old ones, but once you understand it, it’s not that hard to work with. Please refer to the links at the end of this paper which cover the basics on loading images into the database and managing them. It shows how easy it is to do.

Myth 7: Files are best managed by the file system.

Response: Simply, file systems do not scale and have restric-tions associated with them. The security on them does not integrate well with a database. Databases use roles and can be granular. You can’t restrict access to subsets of a file, but if you store that file in a database and separate it out, you can limit it. If you strip out the metadata embedded in it, you can apply roles to it and you can enforce transactional management on the core components. You can tightly secure and audit it. You can recover and restore it.

The challenge that should be made when discussing digital images and storing them in the database shouldn’t be whether one stores them in the database. The starting premise should be we store them in the database. Now justify why you think they should be stored in the file system.

So I hope by reading this article you realise that a lot of what was once thought about unstructured data is based on outdated and now faulty premises. There is a whole world out there, filled with interesting, fun and fascinating data. What are needed are devel-


opers and database administrators who are willing to take on the challenge and attempt to describe and understand all of this new data. What is truly needed are more people willing to take on the challenge of dealing with this data and writing new and innovative papers on how to manage, load, secure and scale it. It’s time to start embracing this data.

For further information see:Unstructured data SIG Site:https://sites.google.com/site/ommuds/

Multimedia BLOG site: http://eternal-donut.blogspot.com.au/

Managing Multimedia in Oracle Book available at: http://www.packtpub.com/managing-multimedia-and-unstructured-data-in-oracle-database/book

i. http://www.oracle.com/technetwork/topics/bigdata/learnmore/online-forum-qa-1536383.html

ii. http://www.autonomy.com/content/Technology/what-is-big-data/index.en.html http://en.wikipedia.org/wiki/Unstructured_dataiii. http://www.digimarc.com/

Marcelle Kratochvil is the Oracle ACE Director for all data/any data. With over 25 years experience as an Oracle DBA, she has spent the last 15 years working extensively on using multimedia in an Oracle database. Her first book on managing unstructured data in Oracle was released in March 2013. She also manages the Oracle Multimedia SIG and has been presenting at OpenWorld on this topic for the last six years. She has worked as a DBA to over 50 sites around the world and has extensive experience and practical knowledge in this field, working on multimedia sites that scale from the very small to the multi-terabyte in size.


the set of historical data in the Foundation table. Beyond that, the view is effectively the same. Now a combination of the indica-tors can be used to return various datasets, including all historical transactions from the initial start date of the data warehouse.

In Summary…As many organizations look to deliver real-time reporting capa-

bilities to their end users, Oracle has a solution available within their stack of data integration products. Oracle GoldenGate, a pre-mier data replication technology, and Oracle Data Integrator, a top-tier data integration tool, are a perfect match and can be combined to deliver real-time, transformed data, to the target data warehouse tables for use in near real-time reporting. These, along with the Oracle Reference Architecture for Information Management, will set you up for a successful real-time data warehouse implementa-tion.

Michael Rainey is a Principal Consultant with Rittman Mead, a leading con-sultancy focused on delivering Oracle Data Warehousing and Business Intelligence technology solutions. Specializing in Oracle Data Integrator, Michael has developed several solutions for migrating various data warehouse tools to ODI using the ODI SDK. In his work, Michael has completed numerous ODI and Oracle GoldenGate implementation projects, sharing his data integration knowledge along the way on the Rittman Mead blog and presenting at conferences such as RMOUG Training Days 2013 and KScope13.

Website: http://www.rittmanmead.comTwitter: @mRainey

LinkedIn: http://www.linkedin.com/in/rainey

Michael Rainey, Golden Gate & Oracle Integrator...continued from page 9

Name RMOUG’s Training Days 2014 Mascot!

http://www.rmoug.org/training/owl-contest/


by Suzanne StrasserDatabase Administrator, Return Path

OverviewOracle Enterprise Manager (OEM) 12c provides a rich set of

metrics for monitoring database health and alerting when problems are encountered. OEM stores data collected at regular intervals for the metrics of interest into a management repository. If we develop an understanding of the structure of the OEM management reposi-tory views, we will gain the ability to extract historical metrics data from the repository to create customized graphs and perform additional analyses outside of OEM.

This article will demonstrate how to:Build queries using the OEM management repository •views to extract metric data for historical analysis, trend-ing, and graphing; andIdentify how OEM uses rollup tables to summa-•rize metric data on an hourly and daily basis.

Extracting Monitoring Data from OEMAt Return Path, we have a business requirement to monitor

the health of our Oracle production databases and servers by col-lecting metrics data on a periodic basis, analyzing the data, and presenting the data using a set of graphs. On a regular basis, we extract the metrics data of interest stored in the OEM management repository and import it into an external graphing tool (such as Excel or Graphite) for historical analysis, trending and forecasting. For this article we will focus on a single metric: CPU utilization on the production database servers. Is the CPU utilization increasing over time? Can we detect any trends in the data?

To answer these questions, we first need to understand the OEM management repository view structures that will allow us to construct the custom SQL queries needed to extract the data. We already have the capability of viewing graphs for each metric within OEM, and they are very useful as a starting point to monitor the CPU utilization, but forecasting and trending capabilities are not built into OEM’s graphing tool. In addition, “out of the box” func-tionality for OEM’s “CPU Usage” category of metrics is set up for real time alerting only, and not for alerting and historical trending, so we will need to make changes for monitoring the host to allow data for this metric to be saved into the management repository for later retrieval.

Another “gotcha” for our business is that one of the produc-tion database servers has only half (16) of its 32 CPUs dedicated to Oracle. The “CPU Utilization %” metric within OEM’s “Load” category of metrics for host target types will report on the CPU usage allocated to all 32 CPUs and not just the 16 dedicated to Oracle. This metric has a default warning threshold of 80% and a

critical threshold of 95%. But since all 32 CPUs are considered, the average percent CPU utilization is calculated across all 32 CPUs and includes the zero values for the CPUs that are not dedicated to Oracle. The end result is an average usage value that is only half of what real usage should be (divide the usage by 16 CPUs and not 32). In order to enable meaningful thresholds for OEM’s “CPU Utilization %” metric, we would need to divide the threshold values by 2 and use 40% and 47.5%.

Also, we want to be able to compare on the same graph the CPU utilization for two different production servers. Only the first server (named db33.lan.returnpath.net) has half of its CPUs being used; the second server (named db103.lan.returnpath.net) has 16 total CPUs and all are dedicated to Oracle. In order to compare apples to apples, we will enable OEM metrics monitoring at the individual CPU level, using the “CPU Utilization % of a CPU” met-ric within OEM’s “CPU Usage” category of metrics for host target types. In addition, for this single metric we will set the collection schedule to every 1 minute in order to store the most detailed level of data into the management repository.

Setting the Host Monitoring ThresholdsWithin the OEM 12c user interface, we select the “db33.lan.

returnpath.net” host target and then select the Host, Monitoring, Metric and Collection Settings menu option. Viewing only the met-rics with thresholds, we change the Warning and Critical thresholds for the “CPU Utilization %” metric from 80 and 95 to 40 and 47.5.

Next, viewing all metrics, we go to the “CPU Usage” group and change the collection schedule to every 1 minute, and set the “Use of Metric Data” to “Alerting and Historical Trending.” This will direct OEM to start collecting values for the metric (all 32 indi-vidual CPUs for db33.lan.returnpath.net) and storing the data into the management repository. For db103.lan.returnpath.net, we will enable the same collection settings but leave the thresholds at their default levels of 80% and 95%.

Querying Metrics Data Using the OEM 12c Management Repository Views


Viewing the Graphs within OEMAfter allowing some time for OEM to collect data into the

management repository, we can view the graphs for each host. We select the db33.lan.returnpath.net host and then select the Host, Monitoring, All Metrics menu option. We expand the “CPU Usage” metric category and then select the “CPU Utilization (%) of a cpu” metric. The right side of the screen will display each individual CPU and the average, low, and high values for various time periods that can be selected from the “View Data” drop down box. Note in the next screen shot that only half of the CPUs are being used, as iden-tified earlier in this article. We select CPU 17 and change the View Data period to the Last 31 Days to see the resulting graph.

Also note that there is missing data for days 23 through 25. This is due to the OEM agent being down on the host server. The agent only collects data real time while it is up, so if your business requires 24x7 monitoring of your database servers, you should explore high availability options for the Oracle Management Service and repository. Also make sure that OEM’s alerts for when an agent is down are emailed to the appropriate on-call personnel that can take care of bringing up the agent.

Understanding the Management Repository Rollup Tables

The previous screen shot shows various levels of detail for the data collected by OEM. This is due to the rollup table strategy implemented by OEM to save space in the repository. The newest data collected, the raw data, is stored at its most detailed level for 7 days in the SYSMAN.EM_METRIC_VALUES table. The raw data is periodically aggregated into hourly averages into the SYSMAN.

EM_METRIC_VALUES_HOURLY table, which is retained for 32 days. The hourly average data is periodically aggregated into daily averages in the SYSMAN.EM_METRIC_VALUES_DAILY table. The daily average data is retained for 12 months before it is purged from the repository.

The Oracle Enterprise Manager Cloud Control Administrator’s Guide describes the data retention policies in Chapter 20. These settings are displayed in the following table:

Table Name Partitions Retained

Partition Size

EM_METRIC_VALUES 7 DAY

EM_METRIC_VALUES_HOURLY

32 DAY

EM_METRIC_VALUES_DAILY

12 MONTH

OEM also provides a query to check the current values of the retention levels:SELECT table_name, partitions_retained FROM sysman.em_int_partitioned_tables WHERE table_name IN (‘EM_METRIC_VALUES’, ‘EM_METRIC_VALUES_HOURLY’, ‘EM_METRIC_VALUES_DAILY’);

OEM also provides a Pl/SQL procedure to allow you to change the data retention levels. Simply identify which of the three tables you wish to change and the new number of partitions to be retained. For example, to change the retention period of the daily averages from 1 year to 2 years, you would execute the following:execute gc_interval_partition_mgr.set_retention(‘SYSMAN’, ‘EM_METRIC_VALUES_DAILY’, 24);

In the above query, ‘SYSMAN’ indicates the owner of the OEM management repository, ‘EM_METRIC_VALUES_DAILY’ is the name of the rollup table to be changed, and 24 is the new number of partitions to be retained (since the partition size for the daily table is in months, the value specifies the number of months of data to be retained).

OEM uses a database scheduled job named EM_ROLLUP_SCHED_JOB to aggregate the raw data into the hourly and daily rollup tables on a nightly basis. By default this job runs daily at 3:00 am. To change the time of the daily run, you could alter the database job to specify a different repeat interval for the schedule. For example, to change the time to execute daily at 4:00 am, you could execute the following Pl/SQL code block:BEGIN SYS.DBMS_SCHEDULER.SET_ATTRIBUTE ( name => ‘SYSMAN.EM_ROLLUP_SCHED_JOB’ ,attribute => ‘REPEAT_INTERVAL’ ,value => ‘FREQ = DAILY; INTERVAL = 1; BYHOUR = 4’);END;/

Understanding the Management Repository Views

To extract the CPU usage data from OEM, we want to be able to construct a query that is appropriate for the level of detail retained in the data. Oracle recommends that we query the data


Using this information we can construct the raw data query, obtaining one row for each minute, and averaging the CPU utiliza-tion across the 16 CPUs: SELECT SUBSTR(entity_name, 1, INSTR(entity_name, ‘.’) - 1) AS db_server, TO_CHAR(TRUNC(collection_time, ‘MI’), ‘YYYY-MM-DD HH24:MI:SS’) AS metric_date, ROUND(MIN(VALUE), 1) AS min_cpu, ROUND(AVG(VALUE), 1) AS avg_cpu, ROUND(MAX(VALUE), 1) AS max_cpu FROM sysman.gc$metric_values WHERE metric_group_name = ‘CPUUsage’ AND metric_column_name = ‘cpuUtil’ AND ( (entity_name LIKE ‘%db33%’ AND key_part_1 BETWEEN 17 AND 32) OR (entity_name LIKE ‘%db103%’))GROUP BY entity_name, TO_CHAR(TRUNC(collection_time, ‘MI’), ‘YYYY-MM-DD HH24:MI:SS’)ORDER BY 1, 2;

Sample output:

DB_

SERVER

METRIC_DATE MIN_CPU AVG_CPU MAX_

CPU

db103 2013-01-06 00:00 21.50 26.10 30.30

db103 2013-01-06 01:00 13.90 17.60 27.30

db103 2013-01-06 02:00 9.40 14.80 22.30

db103 2013-01-06 03:00 9.10 14.90 23.00

db103 2013-01-06 04:00 7.40 13.70 34.20

Hourly Average Data (GC$METRIC_VALUES_HOURLY View)The GC$METRIC_VALUES_HOURLY view corresponds to

the data in the EM_METRIC_VALUES_HOURLY table. The key columns are similar to the columns in the GC$METRIC_VALUES view, but VALUE has been replaced with:

AVG_VALUE – average value of metric for the hour;•MIN_VALUE – minimum value of metric for the hour;•MAX_VALUE – maximum value of metric for the hour; •andSTDDEV_VALUE – standard devia-•tion value of metric for the hour.

COLLECTION_TIME is used to store the date and hour of aggregation.

The most recent hourly aggregate data is for yesterday (due to the time that Oracle rolls up the data). To obtain hourly average data for the current day as well, we will need to union the aggregate data with the raw data from GC$METRIC_VALUES and aggregate it hourly: SELECT SUBSTR(entity_name, 1, INSTR(entity_name, ‘.’) - 1) AS db_server, TO_CHAR(TRUNC(collection_time, ‘HH24’), ‘YYYY-MM-DD HH24:MI:SS’) AS metric_date, ROUND(MIN(min_value), 1) AS min_cpu, ROUND(AVG(avg_value), 1) AS avg_cpu, ROUND(MAX(max_value), 1) AS max_cpu FROM sysman.gc$metric_values_hourly WHERE metric_group_name = ‘CPUUsage’ AND metric_column_name = ‘cpuUtil’

from management repository views stored in the SYSMAN schema instead of directly from the rollup tables. Chapter 18 of the Oracle Enterprise Manager Cloud Control Extensibility Programmer’s Reference explains each of the views and their structures.

To query the CPU utilization for both of our db33 and db103 servers, based on the retention levels we can:

Query the most recent 7 days of raw data (configured •to collect at 1-minute intervals) from the GC$METRIC_VALUES view;Query the most recent 32 days of hourly average CPU uti-•lization data from the GC$METRIC_VALUES_HOURLY view, and union with the most recent raw data from the GC$METRIC_VALUES view (and aggregate the raw data to hourly); andQuery the most recent 12 months of daily average CPU utiliza-•tion data from the GC$METRIC_VALUES_DAILY view, and union with the most recent raw data from the GC$METRIC_VALUES view (and aggregate the raw data to daily).

The next sections show the queries we used to extract the CPU utilization data for each of the above levels of aggregation. Each of the queries selects only the 16 CPUs dedicated to Oracle from db33.lan.returnpath.net and all 16 CPUs in use from db103.lan.return-path.net. The key_part_1 column within each query is used to select only the 16 CPUs in use from db33.lan.returnpath.net.

Raw Data (GC$METRIC_VALUES View)The GC$METRIC_VALUES view corresponds to the data in

the EM_METRIC_VALUES table. The key columns in the view are:

ENTITY_NAME – the target or component name;•COLLECTION_TIME – data collection time in the target •time zone;METRIC_GROUP_NAME – name of the metric group;•METRIC_COLUMN_NAME – name of the metric column;•KEY_PART_1 – key part 1 of composite key; and•VALUE – value of metric. •

The values for the METRIC_GROUP_NAME and METRIC_COLUMN_NAME columns in the view can be obtained by query-ing from the GC$METRIC_COLUMNS view. The values of the METRIC_GROUP_LABEL and METRIC_COLUMN_LABEL col-umns in that view will match the names of the metrics displayed in OEM. In our example, the metric group label is “CPU Usage” and the metric column label is “CPU Utilization (%) of a cpu.” Therefore, we can construct the following query to obtain the metric group name and metric column name:SELECT entity_type, metric_group_name, metric_column_name, metric_group_label FROM sysman.gc$metric_columns WHERE metric_column_label = ‘CPU Utilization (%) of a cpu’;

ENTITY_TYPE METRIC_

GROUP_

NAME

METRIC_

COLUMN_

NAME

METRIC_

GROUP_

LABEL

host CPUUsage cpuUtil CPU Usage


HH24:MI:SS’) AS metric_date, ROUND(MIN(VALUE), 1) AS min_cpu, ROUND(AVG(VALUE), 1) AS avg_cpu, ROUND(MAX(VALUE), 1) AS max_cpu FROM sysman.gc$metric_values WHERE metric_group_name = ‘CPUUsage’ AND metric_column_name = ‘cpuUtil’ AND ( (entity_name LIKE ‘%db33%’ AND key_part_1 BETWEEN 17 AND 32) OR (entity_name LIKE ‘%db103%’))GROUP BY entity_name, TO_CHAR(TRUNC(collection_time), ‘YYYY-MM-DD HH24:MI:SS’)ORDER BY 1, 2;

Graphing the Results Outside of OEMWe used a SQL query tool to extract the data from each of

the queries listed in the previous section to comma-delimited files that could be imported into a Microsoft Excel spreadsheet. In the next series of screen shots, we show some of the graphs that we constructed from the data. We used Excel’s capabilities of adding a linear trend line to determine if there was any increasing trend in the CPU utilization.

Note that the above graph for the raw data has the CPU usage values jumping all over the place for the 1-minute level of aggrega-tion. This is not too useful for creating a trend line but can be used to drill down to view the usage in detail for a portion of the day. In the preceding graph, we have drilled down to show only 24 hours of data.

Further drill down to display only four hours data from 8:00 am to 12:00 pm shows that the CPU utilization tends to spike at the start of each hour. This could be partially explained by many hourly database jobs scheduled to run at the start of each hour. The 8:00 hour contains particularly heavy usage in this example. In this case, if some of the database jobs could be modified to start at times other than at the start of each hour, we might see less spikes in the CPU utilization since the load would be more evenly spread throughout the hour.

AND ( (entity_name LIKE ‘%db33%’ AND key_part_1 BETWEEN 17 AND 32) OR (entity_name LIKE ‘%db103%’))GROUP BY entity_name, TO_CHAR(TRUNC(collection_time, ‘HH24’), ‘YYYY-MM-DD HH24:MI:SS’)UNION SELECT SUBSTR(entity_name, 1, INSTR(entity_name, ‘.’) - 1) AS db_server, TO_CHAR(TRUNC(collection_time, ‘HH24’), ‘YYYY-MM-DD HH24:MI:SS’) AS metric_date, ROUND(MIN(VALUE), 1) AS min_cpu, ROUND(AVG(VALUE), 1) AS avg_cpu, ROUND(MAX(VALUE), 1) AS max_cpu FROM sysman.gc$metric_values WHERE metric_group_name = ‘CPUUsage’ AND metric_column_name = ‘cpuUtil’ AND ( (entity_name LIKE ‘%db33%’ AND key_part_1 BETWEEN 17 AND 32) OR (entity_name LIKE ‘%db103%’))GROUP BY entity_name, TO_CHAR(TRUNC(collection_time, ‘HH24’), ‘YYYY-MM-DD HH24:MI:SS’)ORDER BY 1, 2;

Daily Average Data (GC$METRIC_VALUES_DAILY View)The GC$METRIC_VALUES_DAILY view corresponds to the

data in the EM_METRIC_VALUES_DAILY table. The key columns are similar to the columns in the GC$METRIC_VALUES view, but VALUE has been replaced with:

AVG_VALUE – average value of metric for the day;•MIN_VALUE – minimum value of metric for the day;•MAX_VALUE – maximum value of metric for the day; and•STDDEV_VALUE – standard devia-•tion value of metric for the day.

COLLECTION_TIME is used to store the date of aggregation.The most recent daily aggregate data is for yesterday (due to

the time that Oracle rolls up the data). To obtain daily average data for the current day as well, we will need to union the aggregate data with the raw data from GC$METRIC_VALUES (and aggregate it daily): SELECT SUBSTR(entity_name, 1, INSTR(entity_name, ‘.’) - 1) AS db_server, TO_CHAR(TRUNC(collection_time), ‘YYYY-MM-DD HH24:MI:SS’) AS metric_date, ROUND(MIN(min_value), 1) AS min_cpu, ROUND(AVG(avg_value), 1) AS avg_cpu, ROUND(MAX(max_value), 1) AS max_cpu FROM sysman.gc$metric_values_daily WHERE metric_group_name = ‘CPUUsage’ AND metric_column_name = ‘cpuUtil’ AND ( (entity_name LIKE ‘%db33%’ AND key_part_1 BETWEEN 17 AND 32) OR (entity_name LIKE ‘%db103%’))GROUP BY entity_name, TO_CHAR(TRUNC(collection_time), ‘YYYY-MM-DD HH24:MI:SS’)UNION SELECT SUBSTR(entity_name, 1, INSTR(entity_name, ‘.’) - 1) AS db_server, TO_CHAR(TRUNC(collection_time), ‘YYYY-MM-DD


http://docs.oracle.com/cd/E24628_01/doc.121/e25161/views.htm#sthref1292Oracle Enterprise Manager Cloud Control Administrator’s Guide (Chapter 20):http://docs.oracle.com/cd/E24628_01/doc.121/e24473/toc.htm

About the AuthorSuzanne Strasser is an Oracle database administrator at Return Path, where she

is responsible for management and performance tuning of very large Oracle databas-es in a 24x7 production environment. She has been working with Oracle databases since 1994, with areas of expertise in database performance tuning, monitoring, data modeling, and replication. Suzanne enjoys mentoring software engineers at Return Path on database schema design and SQL query tuning. She recently presented at RMOUG Training Days 2013 and gave a webinar for the Independent Oracle Users Group (IOUG).

About Return PathReturn Path is the worldwide leader in email intelligence. We analyze more

data about email than anyone else in the world and use that data to power products that ensure that only emails people want and expect reach the inbox. Our industry-leading email intelligence solutions utilize the world’s most com-prehensive set of data to maximize the performance and accountability of email, build trust across the entire email ecosystem and protect users from spam and other abuse. We help businesses build better relationships with their customers and improve their email ROI, and we help ISPs and other mailbox providers enhance net-work performance and drive customer retention. http://www.returnpath.com/

The linear trend line for the hourly graph (past 32 days of data) shows a decreasing trend, but the correlation coefficient (R-squared value) is only 0.0041, indicating that only a tiny amount of the data variation can be explained by the trend line.

The linear trend line for the daily graph (past 4 months of data) shows an increasing trend, but the correlation coefficient (R-squared value) is only 0.217, indicating that only a minimal amount of the data variation can be explained by the trend line.

ConclusionOEM 12c collects an extensive set of metrics that can be used

to monitor and present data as required for your business needs. If you become familiar with the structures of the OEM management repository views, you will have the flexibility to query the metric data outside of OEM for various purposes.

ReferencesOracle Enterprise Manager Cloud Control Extensibility Programmer’s Reference (Chapter 18):

❖ 303-330-4065 [email protected] http://arisant.com

Arisant focuses on understanding business requirements

to ensure the most effective implementation of the right

solution. Our strategy is aimed at providing an honest and

expert brand of consulting services for both the Private

and Public Sectors. Contact us today to discuss any

of the following services, and more.

Focused on Delivering the right Solution for the right Project

❖ Consulting Services ◆ Architecture ◆ Analysis ◆ Design ◆ Implementation ◆ Project Management

❖ Staffing/Support

❖ Managed Services

❖ Managed Hosting Services ❖ Oracle Software Sales ❖ Oracle Hardware Sales


I have been working with Oracle’s BI tools for years. I am quite the Discoverer expert (a free tool now from Oracle Corp…OBISE …standard edition) but this tool is now on the back burner and I hardly even field questions on it anymore.

This is part 3 of a 3 part series on getting OBIEE setup (part 1) and creating some reports (part2). This final chapter will expose you to BI Publisher.

You can try these reports out yourself at: http://www.vlamis.com/testdrive-registration/ . This cloud image is the Sample207 image discussed in Part 1. So, you can use the cloud environment provided by Vlamis.com or you can use the same walk thru against the download OBIEE environment discussed in Part 1.

Additional walk thru’s and useful documentation can be found at: http://www.oracle.com/technetwork/middleware/bi-enterprise-edition/overview/index.html

Logging into the Sample207 environment:

http://<ip address>:7001/analytics •login: weblogicpassword: Admin123

The Sample v307 is Oracle BI 11.1.1.7 preinstalled and ready to go is now available. This download is a complete working OBIEE environment for VMBox: http://www.oracle.com/technetwork/mid-dleware/bi-foundation/obiee-samples-167534.html.

Oracle BI Publisher is an enterprise reporting product that provides the ability to create and manage highly formatted reports from a an equally wide variety of data sources. BI Publisher is designed to make use of MS Word and Adobe Acrobat. Other data sources are the Oracle database (based on queries), web services, and Discoverer.

OTN has a nice overview of BI Publisher. Check out this link: http://www.oracle.com/technetwork/middleware/bi-publisher/overview/index.html

BI Publisher is imbedded in OBIEE and is OBIEE’s standard reporting feature. In this article, we will take a report we created in Part 2 and pretty it up using BI Publisher.

I called it ‘Product_Sales’, its properties looks like this:

The data looks like this:

Now, lets publish the report using the built-in BI Publisher in OBIEE. My report Product_Sales is stored in ‘My Folders’ in the browser catalog.

Using BI PublisherLets create a new report. Click New → Data Model

Oracle BI Extended Edition (OBIEE)Tipe & Techniques • By Dan Hotka


Change the Default Data Source to Oracle BI EE…Click on Data Sets and select Oracle BI Analysis from the New

Data Set drop-down menu.

Notice the different supported data sources. Basically, BI Publisher will pull the query from these various sources, or, just format the data from the XML or such…

This will bring up the ‘Create Data Set’ dialog box. Give your report a name and select the magnifying glass next to Oracle BI Analysis…this brings up a Oracle BI Catalog popup. Find your Analysis (mine was under ‘User Folders’ and remember, its Product_Sales.

Your dataset should look something like this. Notice the left column of the data model…you can add events under triggers, you

can add additional fields, list of values, parameters, and bursting is how the report is to be distributed and in which formats.

Once you get a data set, from any of the sources, now it is a matter of what do you want the output to look like. Lets continue.

Up on the upper right…there is now an XML button. Click this to try your query.

You should get something that looks like the above. Click on the ‘Options’ button next to the ‘Return’ button and save the data as ‘sample data’. Now click on the Return button on the upper right corner to return to the data model.

Now that we have a data model, we will now create a report to use it. Remember, this article is a very high and simple use of BI Publisher. BI Publisher can do so many things.

Click on Catalog → My Folders → New → Report

Then choose ‘use existing data model’ then select the data model from the last exercise.


Click on the ‘Interactive Preview’ and select an output for your report. I selected PDF.

Click again on the Interactive Preview button and you should see this:

Now, we have a report that we can select from this interface. We can add the report to a click point on the dashboard. This report is now available thru the OBIEE interface. It can be included in dashboards, printed, scheduled to run, have alerts posted to it (announce when certain data items appear in the report…), etc. The BI Publisher tool can now be used to ‘burst’ or run and distrib-ute the results. The bursting process can produce multiple output types and send to multiple distribution lists, etc.

This is quite the flexible report-handling tool.The latest release of BI Publisher is 11.1.1.7. Some of its new

features include:Can now make PDF’s of OBIEE Dashboards•Retry limit increased when the scheduler fails to start the •jobNow reports against BI Subject areas•Additional charting abilities•Additional CVS import support•ODBC and JDBC support for other data stores•Excel templates made easier•Additional check printing features•Support for ENDECA•Charts can be enhanced easier now•

Let me know how I can help you with your OBIEE implemen-tation.

Dan Hotka is a Training Specialist and an Oracle ACE Director who has over 35 years in the computer industry, over 29 years of experience with Oracle products. His experience with the Oracle RDBMS dates back to the Oracle V4.0 days. Dan enjoys sharing his knowledge of the Oracle RDBMS. Dan is well-published with 12 Oracle books and well over 200 printed arti-cles. He is frequently published in Oracle trade journals, regularly blogs, and speaks at Oracle conferences and user groups around the world.

Dan Hotka - Author/Instructor/Oracle ACE Director

[email protected] 771-3935

Select ‘Report Editor’, give your report a name (I used ‘Product_Sales_Report2’…so it doesn’t conflict with the data model…I should have used a ‘data model’ name in my data model!).

From here, you can select a variety of formats or you can build your own. We will select the ‘Blank Portrait’ report. It is easy to drop back in here and pick a different template. BI Publisher allows you to create these report templates so that all your reports have the same company logo, colors, and general format. How cool is that?

From here, you are ‘report writing’!!! Drag and drop the col-umns into the report. I’ve already created an article that is way too long…but you can do a lot from this editor screen.

Drag and drop your columns from the Data Source onto the canvas.


So what will this do? This will help get everyone familiar with the model and the layout of the diagram, as they have to really look at the diagram to find the words.

For more fun - form teams and keep score! You might even add a time limit per word so they don’t have too much time to study the diagram between rounds.

Eval Commission Job

Salary Department Country ID

Region Postal Code Start Date

End Date Job Employee

Silly SentencesThis idea came from a game in my son’s National Geo Kids

magazine. If you don’t know how this works, you start with sen-tences with blanks in strategic areas. So the sentences may be missing nouns, verbs, adverbs, etc. You have someone fill in the blanks out of context - you ask for a noun but they have no idea what the sentence looks like until after you fill in all the blanks. It can be quite funny.

So how does this apply in data modeling?One of the hardest parts of a logical model is naming the rela-

tionships. Use this game to figure out the right sentences.Start by writing the relationships with completely silly or even

wrong verbs. For example:Each Customer must be found squatting

at one or more Addresses.Each Employee may be hoping for

one and only one Job.

AbstractMost people think data modeling is booooorrring, right? But

for a few of us, it is kind of fun. While data architects the world over will all agree that data modeling is a critical success factor to any well engineered database or data warehouse system, many struggle with how to get their organizations to support their efforts. What if you could make data modeling sessions more engaging for the business folks (your customers!)? If you could, the end result would be better data models. Using some common children’s games and other concepts (and creative license on the themes) this session will show you how to make data modeling fun. Be prepared to rethink how you develop data models.

IntroductionWhile on our annual summer family vacation, I happened to

mention I needed ideas for a blog post. My son, all of nine years old, suggested the above title.

Hmmm...I said...not bad. That might work.After all most people think data modeling is booooorrring,

right?But for a few of us, it is kind of fun.So then I asked him if he had any ideas how we could make

it fun. He did.My son does not actually know how to do any data modeling

(yet), but he has looked over my shoulder a few times and knows I draw pictures with boxes and connecting lines and words in the boxes.

With that bit of knowledge, he did come up with a few good ideas that really could make data model review sessions a bit more fun, and maybe more effective.

The five we ultimately came up with are:Word Search1. Silly Sentences2. Jeopardy3. Data Model Haiku4. Data Model Telephone5.

Word SearchTo do this on you need to put up a large version of a data model

on the wall (you do that already right?). Give the reviewers a list of words to find on the model diagram. You should be able to produce the list using entity and attribute names from your data modeling tool or other data dictionary.

Next have them take turns going up to the diagram to high-light or circle the words on their list.

Five Ways to Make Data Modeling Fun

by Kent Graziano, Data Warrior LLC

Figure 1 - Find the words in the diagram


Each Department must be coloring one or more Job Histories

Use your creativity to come up with goofy verbs for the rela-tionships. Then get the users to “validate” the sentences. In the process you will get them to really think about what the relation-ships should really be.

I am sure they will be more than willing to correct your errors

JeopardyYou all know how this game works - you get the answer and

have to come up with the questions.This is an interesting way to validate your entity and attri-

bute definitions. Use entity definitions as the answers. Users have to guess the entity name.

Consider using definitions that have already been approved to see if they even remember what they told you!

For example: Answer: Person or organization to which

we provide a product or service.Question: What is a customer?

Answer: The primary organizational unit in the company

Question: What is a department?

Of course it will be really interesting to see if they can link defi-nitions you got from them with the entity names in the model. You might get some clarifications in the process.

Another possibility is to use this method as a means of review-ing what you have collected, validating it with the users, then get-ting a sign off.

Data Model HaikuYou can apply this technique with definitions or maybe rela-

tionship sentences. Trying to put the words in a specific form will make you really think about your understanding of the concepts (and force you to be succinct).

In case you don’t remember how to do a haiku poem, it is five syllables, then seven, then five.

So a (boring) example based on a common data model relation-ship is:

Each customer mayBe contacted by one or

More customer reps

In order to make your sentence fit this form, you may have to come up with different words (with the right number of syllables) to express the idea or rule. This may be quite hard but it will test whether or not you really understand the idea. And it will help make sure that the users are clear on their thinking as well.

Note for my friends in the UK: Feel free to do Sonnets in Iambic pentameter.

Data Model TelephoneThis is pretty much what happens anyway - you attend a meet-

ing with the customer, they give you requirements, you take notes,

then try to build a model from those notes. You write out defini-tions and get the users to review those. Chances are good you did not get it quite right.

So for fun, and to make a point about recording details care-fully, and the value of paying attention in a JAD meeting, get your team in a room and start at one end whispering a definition to the first person and have them pass it on. Write down the end result to compare to the definition in the model.

If the result is really funny, tell the customer at the next review meeting.

For your team members who refuse to take notes and say they will remember the definitions well enough to write them into the model later, this could be a very valuable object lesson. Put people like that at the end of the chain.

The truth is, somebody really needs to be a scribe at all JAD sessions. Tough job, but it has to be done. This game can really help convince your team how important that role is. In fact you might even decide it is a good idea to record the sessions to refer to later when you are reviewing the notes.

Consider Your AudienceSo who should you try these with? Most of these can be used

with both your technical team as you do data model reviews and with your business users and SMEs.

The Word Search is a good exercise to use with your technical team to get the ETL programmers, developers, and DBAs familiar with the extent and content of your design.

Silly sentences work well with the business folks as a way to be sure you really understand the business rules and process. It is also a good way to get them to carefully examine their understanding and how they articulate that to the IT team. Likewise Jeopardy is a useful to review and validate all the definitions you have collected and get the sign-off.

Trying to put the relationship sentences and other definitions into haiku form is difficult, but a good exercise for the architect and modelers on the team to make sure they are not getting lazy in pick-ing relationship verbs. If there are a lot of “associated with” and too many “defined by” relationships, then it might be time to try this.

ConclusionSo what do you think? Can we make data modeling more fun? If it helps us to create better data models and gets the business

more engaged in the process, then it is worth a try.If you have any fun ideas, please share!Game on!

About the Author:Kent Graziano is the owner of Data Warrior LLC in The Woodlands, Texas

and a lifetime member of RMOUG and ODTUG. He is a certified Data Vault Master (DVDM), Oracle ACE Director, and expert data modeler and architect with 30 years of experience, including over 20 years working using Oracle, Oracle tools, and doing data warehousing. Kent has written numerous articles and done over 50 presenta-tions (both nationally and internationally). He was the recipient of the 1999 Chris Wooldridge Award (from IOUG) for outstanding contributions to the Oracle user com-munity. In 2003 he was presented with The Doug Faughnan Award for his dedicated service and outstanding contributions to RMOUG. In 2007, he was the recipient of the ODTUG Volunteer Award. He is a co-author of four books including Oracle Designer: A Template for Developing an Enterprise Standards Document, and Supercharge Your Data Warehouse. His most recent publication is A Check List for Doing a Data Model Design Review available in Kindle on Amazon. Kent can be contacted at [email protected] or followed on twitter @KentGraziano. His author page is available here: http://www.amazon.com/Kent-Graziano/e/B000APUYH4/.


by Lena CohenRMOUG Member Focus

Over the years, I feel like a perma-nent student. The moment you

begin feeling comfortable and think you can share a tip or two with your fellow col-leagues, suddenly the landscape changes, and you find yourself in the enchanted forest. It is amazing, but a 20-year-old computer today is perceived as a dinosaur, and nobody would possibly even want to get it as an antique relic. At the same time, a 20-year-old car might even serve your needs by getting you from point A to point B. I still keep wondering how computer technol-ogy would evolve and, most importantly for me, what would be my place in it.

My first encounter with a computer was at the Tel-Aviv university library, and at that time I didn’t anticipate I could use it any way other than checking books out. Only after completing my BS in chemistry, I started to question my interest in that par-ticular subject. My career direction materi-alized into something totally different from the initially set course. Opportunity pre-sented to re-educate myself and step into a fast evolving industry which demanded analytical thinking, and I took it.

I have eagerly studied everything related to databases, programming and the telecommunication industry. Years later, I supported multiple projects around the globe. That time of my life was defined as “sleeping with my suitcase prepared.” Every other month I was flying to other customer site in England, Ireland, Australia, Canada and the USA to set up and support Oracle databases. It was an interesting and chal-lenging time since sometimes you needed to fit into somebody’s big shoes straight away.

“How should we resolve this lock-ing problem?” “Why is my process stuck?” “What should be done to compile processes with the new oracle release? Sometimes, I had no idea how to answer these questions. I had some experience how to research, connect what I heard from others and try to apply it , preferably not on any live sys-

tem. After few “applications” of this sort, I understood the real meaning of the Russian proverb “Measure seven times, cut once”. It was a great way to learn, discover new countries, meet smart people and establish connections. The truth is, I could not keep up with this way of living for long.

This made me to pursue relocation opportunity, which brought me and my husband to a contract position in Ireland. We were working for an Ireland telecom-munication provider. I still remember how difficult it was to obtain 2Gb of space to extend our 200Gb database. I recall hav-ing more spare time on my hands for coffee breaks because all operations involved a great deal of waiting. It took several days to clone a database. Potentially, every little mistake could cost additional time to repeat the process all over again.

One of the exciting events I could recall from that time was the upgrade of Oracle to version 8, and the plan was to do it on a Saturday. One problem we didn’t really account for – there is a very strong Irish tradition to go to the pub on Friday. My first experience for the installation of live production systems happened in an

unexpected way. Only later, I realized I needed to step in with not much of prepara-tion time, because… I was the most sober between all local DBA teams. I carried on this task with my colleague on the phone. The system was not nearly big as modern systems, but still involved several comput-ers and the compilation of all applications. Unbelievable to myself, the full system came up on Monday with no problem.

My work in the USA started at 2001. It’s been a roller coaster with ups and downs and many interesting projects. I have spent many years working for the telcom industry and billing systems. This type of business demanded long hours and great dedication. Meanwhile , due to sensi-tive flows and robust systems, new features and advancements were coming somewhat

slow. So I thought initially that simple partitioning or implementation or patches or new versions needed long months of test-ing. The development and implementation cycle was very long and complicated.

But the landscape changes still came very fast. I remember early discussions about implementation of GPS systems as part of a billing service. It sounded next to impossible at that time. Telecommunication services started to include mobile to origi-nally supported landlines. It feels like a


RMOUG Member Focus

few moments later, and the phone became a computer extension. New and different ways of measuring and categorizing inter-net data had to be developed. Things like interactive games and access internet data have become popular new services that the industry makes money from. Database sizes increased dramatically.

We needed solutions to improve per-formance, so we could sustain the amount of new data. Initially, we were carefully sepa-rating hot objects on different tablespaces which could be potentially separated on the storage level. We tried to utilize the same tablespace for objects which would not be accessed concurrently. This helped only until a change in schedule or applica-tion flow. Nowadays, sophisticated storage systems and ASM make these problems nonexistent.

Another solution to sustain the new load was in data segregation, so we could process logically different data on separate databases and servers. As you can expect, other problems rose while generating con-solidated pictures or during processing data coming from multiple separate server systems. The infrastructure resembled a huge spider web where any small change

of predefined flow could cause a huge performance offset. As I see it today, data consolidation is a new “trend” which would help to simplify and improve database systems. I look back amazed at how things have changed and how different we resolve problems today. Sometimes I even have a tendency to wish things would not change so fast, but at the same time I am excited to explore future innovations.

Several years of my expertise were spent to manage a DBA infrastructure team for an Advertisement and Media project. This project in particular taught me to work in a multicultural environment. It helped me technically and professionally but also taught me to embrace differences in people. Each person has their own per-ceptions whence it could trigger responses different from what you expect. Lessons learned – the natural tendency of mak-ing a virtual catalog of stereotypes and predefining people with labels is far from being a bullet-proof solution. I do always remember that respect is the keystone for successful communication, and it applies to customer relations as well as to great team work. The ideal working environment for me specifically also encourages out of the box thinking where people are expected to propose and evaluate.

Success goes to those who are willing to be different in ways that bring value to others. (Rich Horwath)

The A&M project brought me and my family to Denver, where we have stayed the last nine years. My family expanded here from 2+1 to 2+3. We love and cherish every moment we have spent in Colorado. It’s been a blessing to have wonderful friends and enjoy our environment and nature. Even if sometimes we miss the sea-shore and our family abroad, it is always good to be here.

Another lucky coincidence is that our state is a home for Rocky Mountain Oracle Users Group. I have always wondered if it would be ever possible to create a huge pool of knowledge for techno-logical advancements Everybody who is interested could just plug his brain in and get a quick upgrade, while increasing com-mon resources and capabilities. Today we have exactly that – a big knowledge center called RMOUG, thanks to all the people

who make this engine tick. I admire the Rmoug leading team for the great work and enthusiasm which they share with com-munity. It is great to associate myself with such a great, dedicated and strong team.


A year in review…. When I look back over the last year, I

am amazed at what I have done and have been associated with. Let me introduce myself; I’m Bobby Curtis. I’m a father, husband, database administrator and part-time entrepreneur. I have been in the information technology field since 1996 and worked on just about every platform there is available in the industry. I would be lying if I said I have had fun in the information technology industry for the 17 years. I have been working in it. For the most part, it has been a productive 17 years with quite a lot of ups and downs throughout. My IT career started off in the US Army where I developed skills as a systems administrator supporting networks in the far reaches of the world. After military service, I was a senior systems administrator until I made the transition to being a DBA in 2004 - which brings me back to the opening sen-tence in this article. This last year (09/2012 – 09/2013) has been one of the most exciting and enjoyable years I’ve had in the informa-tion technology industry, both career and personally.

This time last year, I was preparing for my first journey to a major conference within the industry. You can properly guess which conference I was heading to; yep, Oracle Open World 2012! It wasn’t enough that I had the opportunity to got to OOW12, I was also scheduled to speak. What speak at a major conference? I can say I was excit-ed, but in reality I was a bit scared. When I got my speaker slot and it was Thursday the week of the conference; I was a bit upset along with relieved. The upside to speak-ing on the last day was it provided me a chance to watch others present throughout the week and take a few pointers. One such presentation I attended was on Oracle Enterprise Manger 12c Extended Metrics. Going into the presentation, I was like, “Lets see. What this is about?” Needless to say, I was pretty blown away, not only by the topic but the speaker. Unfortunately,


I couldn’t make it to the front of the room to talk with the speaker. I figured I might catch the person somewhere else at the con-ference; I left without speaking to anyone but impressed. Needless to say, I was also paranoid about speaking after seeing that presentation.

The next morning, I was wondering around the OOW grounds and ended up at the demo grounds. I ran into the presenter I wanted to talk with the day before. This person was more than willing to spend some time to talk with me. We talked for about an hour and then parted ways. The insightful knowledge that was discussed I have used since. This person also eased my nerves about present-ing at OOW.

After returning back to Atlanta, I was driven to do more work and expand on my knowledge base of Oracle products and within the com-

munity. Shortly after I got an email from IOUG about the Collaborate conference. I was still riding my high from OOW so I sub-mitted for Collaborate. Early 2013, I was notified that I would be speaking in April at Collaborate. I was excited to speak at my second conference. Shortly after, the speak-er I met at OOW had contacted me about doing beta testing for the new release of the Oracle database. I was excited to accept the opportunity and even more excited that BIAS allowed me to go.

The week in February for beta test-ing was exciting. Getting to work with, test, break, and recommend changes to an Oracle product was not only exciting, but the contact and acquaintances made were worldwide. By the end of the week, users from all the major user groups had made recommendations on what needed to be fixed and either provided approval or provi-sional approval for Oracle’s next generation database.

After getting back from a week of beta testing, I was scheduled to go to Exadata Sales training. So again, I was off jetting to another location for business purposes, as I have done for much of the year. While at training, I had a question around an OEM problem I couldn’t solve. So, I contacted my friend from OOW. They pointed me to the

by Bobby CurtisRMOUG Board Focus

SQL>UPDATE • Fall 2013 27 SQL>UPDATE • Fall 2013 27

RMOUG Board Focus

information I was over-looking. While talk-ing with them, I asked how the book was coming. After discussing the book, I was asked to if I would like to write a few chap-ters in it. First, I was kinda taken back; after thinking about it, I accepted and off writing I went.

As the book was coming together, discussions around ODTUG KScope 13 started. I initially was not going to attend. After discussing it with a few people, I was able to get a speaker spot and present on the beta testing, which I participated in during February. KScope was defiantly the best of the three conferences I did over the last year, not only content, but also in the connections made and the content of the conference.

Now, Oracle Open World 2013 approaches, and I find myself starting where last year began. It is a little unreal to look back at all the items accomplished and the people I met over the last year. I can count this year as one of the years where my career in IT has taken a positive turn and has taught me much more than just

database skills. I would encourage everyone to find a mentor they can trust and a way to get involved with the Oracle community. It is full of people who not only love technol-ogy, but are willing to include you in many different aspects if you allow yourself.

In closing, I have to say, that I couldn’t have done a lot of items this past year without the support of not only BIAS, but my family. My lovely wife, Patty and three wonderful kids, Cole, PJ and Addie, all have given up time to allow me to progress in my career. For that, I’m truly blessed!


Meet Your Board

Tim GormanPresident & Vendors DirectorE-mail: [email protected]

John PetersonVice President

E-mail: [email protected]

Ron BichSecretaryE-mail: [email protected]

Thomas GreenTreasurer


Mark JamesBob Mason-AssistantMember Relations DirectorE-mail: [email protected]

Carolyn FrycVice President & Programs DirectorE-mail: [email protected]

RMOUG Board of Directors

Pat Van BuskirkNewsletter Director


Art MarshallWeb DirectorE-mail: [email protected]

Chris OstrowskiVendor Relations DirectorE-mail: [email protected]

Dan HotkaEducation Director



SQL>UPDATE • Fall 2013 29 SQL>UPDATE • Fall 2013 29

RMOUG Board of Directors

Kellyn Pot’vinTraining Days DirectorE-mail: [email protected]

Bobby CurtisSocial Media DirectorE-mail: [email protected]

Kathy RobbBoard Member Emeritus

Arisant, LLCE-mail: [email protected]

Heidi KuhnExecutive DirectorVoice Mail: (303) 948-1786Fax: (303) 933-6603E-mail: [email protected]

Vince GiasolliSpecial Interest Groups/Meetup DirectorE-mail: [email protected]

Bob MasonMark James-Assistant

Scholarships DirectorE-mail: [email protected]

Kumar AnnamniduMemberAt Large


John JeunnetteBoard Member Emeritus


Peggy KingPast PresidentE-mail: [email protected]


9/22-26/13 Open World Oracle Open World 2013, San Francisco, CA (www.oracle.com/openworld)10/1/13 SQL>Update Call for Articles & Cover Photo Submissions10/29/13 UTOUG Fall Sumposium, Sandy UT (www.utoug.org)11/1/13 SQL>Update Article & Cover Photo Submission Deadline11/15/13 QEW RMOUG Autumn Quarterly Educational Workshop (QEW), Oracle DTC office in Denver CODecember OTD Oracle Technology Day, TBD Location, Denver Area (www.rmoug.org)12/15/13 SQL>Update Publication/Mailing Deadline2/5-7/14 Training Days RMOUG Training Days 2014, Colorado Convention Center in Denver, CO (www.rmoug.org)3/2-6/14 Hotsos Symposium 2014, Omni Mandalay, Iiving TX (www.hotsos.com)4/7-11/14 Collaborate Collaborate 2014, Las Vegas, NV (www.ioug.org)6/22-26/14 ODTUG KScope14, Seattle, WA (www.odtug.com or www.kscope14.com)Please note dates are subject to change. For the most current events calendar visit our website at www.rmoug.org.

RMOUG Events Calendar

Breakfast Discounted Total Cost Ad Rate 1/4 Page $350.00 $175.00 $ 525.001/2 Page $350.00 $312.50 $ 662.50Full Page $350.00 $500.00 $ 850.00Inside Cover $350.00 $625.00 $ 975.00Back Cover $350.00 $750.00 $1,100.00

Contact Carolyn Fryc - Programs Director - 720-221-4432 - [email protected]

Help RMOUG Members and Receive Recognition in An Upcoming Issue of

SQL>Update

SPONSOR A QUARTERLY EDUCATION WORKSHOP AND RECEIVE A HALF

PRICE AD IN SQL>UPDATE

Tell Us About Yourself

Join us in sharing your Oracle experi-ences with other RMOUG members!

Tell us about your life, your job, or share your amusing Oracle anecdotes, tips and secrets!

Please submit all material to [email protected]

Tired of the I-25 Corridor?60 Acres In Gorgeous Bijou Basin

Less Than An Hour From DTC

• Grassy Meadows & Pines• Excavated Road and Building Site• Power & Phone At Property Line• Water Rights• One Hour From DIA• Possible Owner Carry W.A.C.

30777 Magic Dog Circle, Kiowa, CO $269,000

Pat Van Buskirk • Coldwell Banker, Parker Office(303) 243-0737 Cell

www.patvanbuskirk.com [email protected]


Index To AdvertisersArisant .................................................................................. 18Centroid ................................................................................ 12DBAK .................................................................................... 13Quarterly Education Workshop ........................................... 31Quarterly Education Workshop Sponsorship ..................... 30Regis University ................................................................... 2Real Estate Coldwell Banker ............................................... 30RMOUG Name The TD 2014 Mascott ............................... 13Training Days 2014 .............................................................. 32Training Days 2014 Volunteers ........................................... 2

Reach A Targeted Oracle Audience Advertise Now!

A full page, full color ad in RMOUG SQL>UPDATE costs as little as 70 cents per printed magazine and even less for smaller ads.

Contact [email protected]

Join us for our next Quarterly Education Workshop in August for presentations at the Oracle DTC Campus. RMOUG hosts quarterly workshops in May, August and November of each year with the fourth and largest educational event being Training Days in February. Learn about the newest technologies, gain more insight into Oracle techniques and enjoy the camaraderie of meeting with other Oracle professionals.

If you or your organization are interested in partnering with RMOUG to host an upcoming meeting, or to submit an abstract for presentation, please contact

Carolyn Fryc, Programs Director at [email protected]

Watch RMOUG’s Web Page for November Training Topics www.rmoug.org

November 15, 2013Quarterly Education Workshops

suzanne strasser · 2019-11-07 · make business decisions. oracle data integrator 11g and oracle...

Documents