grid resouce management pham thanh toan trinh quoc huy pham the anh 10/18/2015 1
TRANSCRIPT
GRID RESOUCE MANAGEMENTGRID RESOUCE MANAGEMENT
Pham Thanh Toan
Trinh Quoc Huy
Pham The Anh
04/21/231
Outline
Introduction to Grid Resource Management Hierarchical Model Abstract Owner Model Economy/Market Model
204/21/23
Introduction
Resource types
304/21/23
Physical resource: computer, disk, database, networks, scientific instruments.
Logical resource: Executing applications, Complex workflows…
Introduction
Definition
404/21/23
The term resource management refers to the operations used to control how capabilities provided by Grid resources and services are made available to other entities, whether users, applications or services.
Introduction
Definition
504/21/23
At the heart of the Grid is the ability to discover, allocate, and negotiate the use of network accessible capabilities, be they computational services offered by a computer, application services offered by a piece of software, bandwidth delivered on a network, or storage space provided by a storage system.
Introduction
Requirements
604/21/23
Establish a mutual agreement between resource provider and resource consumer.
Encompass a wide range of different scenarios: task submission, workload management, on-demand access, co-scheduling, resource brokering,…
Encompass not only a commitment to perform a task but also commitments to level of performance, or quality of services.
Introduction
Issues
704/21/23
Coordinate resource usage, the diversity of resource types and the variety of different management modes.
The differing policy requirements of resource owners make difficulties in creating a consistent cross-organized policy domain that delivers the necessary capability to the end user while respecting the policy requirements of the resource owner.
Grid Applications require the concurrently allocation of multiple resources.
Introduction
Grid Resource Management Models
804/21/23
Hierarchical model. Abstract Owner Computational Market/Economy Model.
Hierarchical Model
04/21/23 9
Hierarchical Model
Hierarchical Resource Management
The outcome of Grid Forum’s second meeting. GRAM, Legion, Ninf. The major components of this model is divided into
passive and active components
04/21/23 10
Hierarchical Model
Passive components Resource:
things that can be used for a period of time, and may or may not be renewable
They have owners who may charge others for using resources.
They can be shared or exclusive.
04/21/23 11
Hierarchical Model
Passive components Task:
Are consumers of resources Include both traditional tasks and non-computational tasks
(such as file and communication). Jobs:
Are hierarchical entities: can be composed of sub-jobs or tasks, and sub-jobs maybe themselves contain sub-jobs. The leave of this structure are tasks.
The simplest form of a job is one containing a single task
04/21/23 12
Hierarchical Model
Passive components Schedules:
Are mappings of tasks to resource overtime.
04/21/23 13
Hierarchical Model
Active components Schedulers:
Compute one or more schedules for input jobs The unit of scheduling is the job. The schedulers outside control domains cannot commit
resources; these are known as metaschedulers or super schedulers.
Information Services: Act as databases for describing items of interest to the
resource management systems: resource, jobs, schedulers, agents, …
Could be a LDAP, a database,…
04/21/23 14
Hierarchical Model
Active components Domain Control Agents (Local resource manager): Commit resource for use The set of resources controlled by an agent is a control
domain. Are distinct from Schedulers, but control domains may contain
internal Schedulers. Can provide state information, either through publishing in an
Information Service or via direct querying. The actions outside are called requests. Action inside maybe
called command.
04/21/23 15
Hierarchical Model
Active components Deployment agents: implement schedules by negotiating
with domain control agents to obtains resources and start tasks running.
Users: submit job to the resource management system for execution.
Admission Control Agents: determine whether the system can accommodate additional jobs, and reject or postpone job when the system are saturated.
04/21/23 16
Hierarchical Model
Active components Monitors:
Track the progress of jobs. Obtain job status from the task comprising the job and
from the Domain Control Agents where those tasks are running.
Perform outcalls to Job Control Agents and Schedulers to effect remapping, terminating extending lifetime, …
04/21/23 17
Hierarchical Model
Active components Job Control Agents:
Are responsible for shepherding a job through the system, and can act both as a proxy for the user and as a persistent point for a job.
Are responsible for coordinating between different components within the resource management system.
04/21/23 18
Hierarchical Model
Interaction of Components
Job Control Agent
Job Control Agent
UserUserA user submit a job to a Job Control Agent.
04/21/23 19
Hierarchical Model
Interaction of Components
Job Control Agent
Job Control Agent
Admission Control Agent
Admission Control Agent
UserUser
Job Control Agent calls an Admission Agent to examines the resource demands of the job (perhaps consulting with a grid Information System).
04/21/23 20
Hierarchical Model
Interaction of Components
Job Control Agent
Job Control Agent
Admission Control Agent
Admission Control Agent
UserUser
SchedulerScheduler
The Admission Agent pass the job to the Scheduler, which performs resource discovery (using the grid Information System).
04/21/23 21
Hierarchical Model
Interaction of Components
Job Control Agent
Job Control Agent
Admission Control Agent
Admission Control Agent
UserUser
SchedulerSchedulerDomain Control Agent
Domain Control Agent
Scheduler consults with Domain Control Agent to determine the current state and availability of resources.
04/21/23 22
Hierarchical Model
Interaction of Components
Job Control Agent
Job Control Agent
Admission Control Agent
Admission Control Agent
UserUser
SchedulerSchedulerDomain Control Agent
Domain Control AgentDeploymentDeployment
The Scheduler computes a set of mappings and passes these mappings to a Deployment Agent.
04/21/23 23
Hierarchical Model
Interaction of Components
Job Control Agent
Job Control Agent
Admission Control Agent
Admission Control Agent
UserUser
SchedulerSchedulerDomain Control Agent
Domain Control AgentDeploymentDeployment
The Deployment Agent negotiates with the Domain Control Agents for the resource indicated in schedules and obtains resource reservations for the resources
04/21/23 24
Hierarchical Model
Interaction of Components
Job Control Agent
Job Control Agent
Admission Control Agent
Admission Control Agent
UserUser
SchedulerSchedulerDomain Control Agent
Domain Control AgentDeploymentDeployment
These reservations are passed to Job Control Agent.At the proper time, the Job Control Agent works with different Deployment Agent, and the Deployment Agent works with the appropriate Domain Control Agents to start the tasks running.
04/21/23 25
Hierarchical Model
Interaction of Components
Job Control Agent
Job Control Agent
Admission Control Agent
Admission Control Agent
UserUser
SchedulerSchedulerDomain Control Agent
Domain Control AgentDeploymentDeployment
MonitorMonitor
A Monitor tracks progress of the job and may later decide to reschedule if performance is lower than expected.
04/21/23 26
Abstract Owner Model
04/21/23 27
Abstract Owner Model
Abstract Owner Model Introduction General Structure of AO Grid Resources Negotiating with an AO Job Shops
04/21/23 28
Abstract Owner Model
Abstract Owner When one makes a long distance phone call, who
"owns" the resource being used?
04/21/23 29
Abstract Owner Model
Abstract Owner Each grid resource be represented by one or more
“abstract owners” (AO) For complex resources, an AO will certainly be a broker
for the actual owners or other brokers Pre-existing contract or protocol to acquire resource
04/21/23 30
Abstract Owner Model
General Structure of AO - Resource Resource: any combination of hardware and software
A processor Another Grid system
Resource will be regarded as object Methods: initiating and controlling tasks Attributes: customize the resource
04/21/23 31
Abstract Owner Model
General Structure of AO - Task Task: problem/part of problem that need resource to do.
A message to be sent A process to be executed A complete multi-process program
04/21/23 32
Abstract Owner Model
General Structure of AO–External view Order Window: negotiate to acquire resource
how soon the resource may become available? how much it might cost?
Pickup Window: deliver resource to client
04/21/23 33
Abstract OwnerAbstract Owner
Order Window Pickup Window
External View of AO model
Abstract Owner Model
General Structure of AO – External View
04/21/23 34
Client send request and negotiate to AO
successful negotiation?
Client place order to AO through Order Window
AO return resource to Client through Pickup Window
Client wait for resource?
CANCEL
N
Y
Y
N
Abstract Owner Model
General Structure of AO – Internal View InternalStructure
An owner of a physical resource
04/21/23 35
Resource ManagerResource Manager
Order Window Pickup Window
AO is Resource Owner
Physical ResourcePhysical Resource
Abstract Owner Model
General Structure of AO – Internal View Internal Structure
Higher level broker
04/21/23 36
AO is Broker
ManagerManager
Order Window Pickup Window
Sales Rep.Sales Rep. Delivery Rep.Delivery Rep.
AO3AO3
AO2AO2
AO1AO1
Abstract Owner Model
Grid Resources Three type of resources:
Instruments• Logically exists at some location.
• Creates, consumes, or transforms data or information.
Channels• Exists to facilitate the explicit transfer of data or information between
two or more instruments
• Connects to an Instrument through a Port
Complexes• A collection of (connected) Channel and Instrument resources.
04/21/23 37
Abstract Owner Model
Grid Resources Instruments are divided into subclasses
Compute Instrument: a processor or set of processors along with associated memory, temp files, software, etc.
Archival Instrument: persistent storage of information. Personal Instrument: interface directly to a human being,
ranging from a simple terminal to a more complex
04/21/23 38
Abstract Owner Model
Negotiating with a AO Client need to negotiate with AO before using resource Negotiating process:
Client sends negotiation AO replied with appropriate resource information Client accept and place order, or reject the negotiation.
04/21/23 39
Abstract Owner Model
Negotiating with AO Sample object attributes:
Client first effectively creates a "sample" resource object of the appropriate structure
Assigns each attribute either • (1) a constant value,
• (2) a "don’t care" value,
• (3) a variable name
Some variables can be “negotiable” Variable constraint list: constraints for variables
04/21/23 40
Abstract Owner Model
Negotiating with AO Negotiation Style:
Immediate: AO is to schedule the resource immediately Pending: AO is to return a specified number of sets of
proposed values for the negotiation variables Confirmation: AO is to finish scheduling based on an
earlier-returned set of negotiation variable values Cancel: AO is to cancel earlier Pending negotiation
Pickup Approach: Protocol to be used at Pickup Window
04/21/23 41
Abstract Owner Model
Negotiating with AO Authorization: a capability or key which allows the AO to
determine the authority of the client to access resources Bid: maximum price that the client is willing to pay for the Negotiation ID: like cookie
04/21/23 42
Abstract Owner Model
Job Shop Instead of only deal with Resource acquiring, “Job Shop”
handle job scheduling. Job Shop
Estimator Executor
04/21/23 43
Job Shop(Estimator & Executor)
Job Shop(Estimator & Executor)
Job Result
Job Scheduling
AO for GridAO for Grid
Resource Des. Resource
Abstract Owner Model
Job Shop – The Estimator Deals with the customer, acting as a liaison with the grid
AO (through its order window) if necessary to help determine how soon the job might be done and how much it might cost.
Requests the resources needed from the grid AO (through the order window).
Records (in the Work To Do List) what needs to be done when the resources are ready.
04/21/23 44
Abstract Owner Model
Job shop – The Executor Takes ready resources from the AO delivery window Dequeues the associated work (tasks) from the Work To
Do List Builds any necessary environment for those tasks based
on the resource Initiates the tasks on the resource and collects the
answer. Frees the resource upon completion, if necessary. Notifies and returns the answers to the user
04/21/23 45
Abstract Owner Model
Job Shop
04/21/23 46
EstimatorEstimator
Job Shop
ExecutorExecutorWork to do listWork to do list
Abstract Owner Model
AO Summary AO model is still a proposal There are many remaining gaps in the model Even a potentially useful and well-defined (successfully
prototyped) AO protocol will not be viable unless it can coexist with other contemporary approaches.
04/21/23 47
Economy/Market Model
04/21/23 48
Economy/Market Model
Overview Grid environment
Resources are geography distributed and belong to different organization
Has different resource management, policies, charges for different users
Need a market-driven resource management system to satisfy both resource owner and user
• User : minimize expenses
• Resource owner: maximize return-on-investment
04/21/23 49
Economy/Market Model
Overview Necessitating a grid resource management system:
Provides appropriate tools and services to allow both resource users and owners to express their requirements.
• Users: specify their “QoS requirements” such as minimize the computational cost and yet meet the deadline for the results.
• Resource owners: specify their charges, that can vary from time to time and users to users—and terms of use.
Likely that more and more systems are going to emerge based on this concept.
04/21/23 50
Economy/Market Model
Overview Economy/Market model for grid resource management
Captures the essentials of both • Hierarchical model
• Abstract Owner model
Many of the similar ones appears to be futuristic But points out the need for economy in computation
implicitly.
04/21/23 51
Economy/Market Model
Overview The key components of economy-driven resource
management system: User Applications (sequential, parametric, parallel, or
collaborative applications) Grid Resource Broker (a.k.a., Super/Global/Meta
Scheduler) Grid Middleware Domain Resource Manager (Local Scheduler or Queuing
system)
04/21/23 52
Economy/Market Model
Overview
A possible architectures for grid resource management based on computational market model04/21/23 53
Economy/Market Model
Grid Resource Broker (GRB)
Acts as a mediator between the user and grid resources using middleware services.
Is responsible for Resource discovery. Resource selection. Binding of software (application). Data, and hardware resources,. Initiating computations. Adapting to the changes in grid resources. Presents the grid to the user as a single, unified resource.
04/21/23 54
Economy/Market Model
Grid Resource Broker (GRB)
Includes Job Control Agent (JCA): is a persistent central component
• responsible for shepherding a job through the system.
• takes care of schedule generation,
• the actual creation of jobs,
• maintenance of job status,
• interacting with clients/users, schedule advisor, and dispatcher.
Schedule Advisor (Scheduler): to select those resources that meet user requirements such as meet the deadline and minimize the cost of computation; is responsible for
• resource discovery (using grid explorer),
• resource selection, and job assignment (schedule generation).
• Its key function is while assigning jobs to resources.
04/21/23 55
Economy/Market Model
Grid Resource Broker (GRB)
Includes (cont.) Grid Explorer:
• is responsible for resource discovery by interacting with grid-information server
• identifying the list of authorized machines• keeping track of resource status information.
Trade Manager (TM): • works under the direction of resource selection algorithm (schedule
advisor) to identify resource access costs. • interacts with trade servers and negotiates for access to resources at
low cost. • It can find out access cost through grid information server if owners post
it. Deployment Agent:
• is responsible for activating task execution on the selected resource as per the scheduler’s instruction.
• periodically updates the status of task execution to JCA.04/21/23 56
Economy/Market Model
Grid Middleware Offers services that help in coupling a grid user and (remote)
resources through a resource broker or grid enabled application. It offers core services Remote process management Co-allocation of resources Storage access Information (directory) Security, authentication Quality of Service (QoS), such as resource reservation for
guaranteed availability and trading for minimizing computational cost.
Some of these services have already been discussed in the hierarchical model.
04/21/23 57
Economy/Market Model
Grid Middleware Main components
Trade Server (TS): • It is a resource owner agent • Negotiates with resource users and sells access to resources. • Aims to maximize the resource utility and profit for its owner• Consults pricing algorithms/models defined by the users during negotiation • Directs the accounting system to record resource usage.
Pricing Algorithms/Methods: • Define the prices that resource owners would like to charge users. • The resource owners may follow various policies to maximize profit and
resource utilization and the price they charge may vary from time to time and one user to another user and may also be driven by demand and supply like in the real market environment.
Accounting System: • Responsible for recording resource usage • Bills the user as per the usage agreement between resource broker and trade
server (resource owner agent)04/21/23 58
Economy/Market Model
Domain Resource Manager Local resource managers
Responsible for managing and scheduling computations across local resources such as workstations and clusters.
Even responsible for offering access to storage devices, databases, and special scientific instruments such as a radio telescope.
04/21/23 59
Economy/Market Model
Comments There’re numerous methods for determining/knowing access cost.
The services offered by trade server could also be accessed from or offered by grid information
A trade manager or broker can directly access information services to identify resource access cost and then contact resource agents for confirmation of access.
The trade manager can use these advertised/posted prices or ask/invite for competitive quotes or bids and choose resources that meet user requirements.
The proposed figure is one of the possible alternatives for computational market model and it can vary depending on, particularly, trading protocols like in real world economy.
The different grid systems may follow different and it will be beneficial if they are all interoperable.
The interoperability standards can be evolved through grid user/developer community forums or standardization organizations
04/21/23 60
Economy/Market Model
Q & A
04/21/23 61
Economy/Market Model
References
04/21/23 62
Rajkumar Buyya†, Steve Chapin*, and David DiNucci, Architectural Models for Resource Management in the Grid, USA, 2000
Ian Foster and Carl Kesselman (editors), The Grid: Blueprint for a Future Computing Infrastructure, Morgan Kaufmann Publishers, USA, 1999