Five Things You Need to Know Today About Disaster Recovery PlanningThe Changing Face of Disaster Recovery

Table of ContentsIntroduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3

Multiple Flavors Aren’t Just for Ice Cream Anymore . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4

Virtual This, Virtual That—Let’s Get Real . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5

When Did Weather Metaphors Become So Common? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6

Wait…This Sounds Expensive? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7

Hope Is Not a Strategy — the Hidden Secret of Disaster Recovery Planning . . . . . . . . . . . . . . . . . . . . . . . . . . 8

Fantastic! What’s Next? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9

3

IntroductionAs a busy data center manager, you might be struggling to keep up with the challenge of continually doing more with fewer resources . IT has become an integral part of business operation; in fact, most, if not all of the infrastructure you support is critical for keeping your company up and running . However, despite being asked to do more with less, you probably haven’t gotten any more budget or team members, so you may have been forced to put some projects on the back burner .

Chances are disaster recovery is one of those projects you’d like to bring to the front of the priority queue, and for good reason . It’s nearly impossible to turn on the news without being bombarded by all the potential risk out in the world . Whether it is turbulent weather, natural disasters or man-made accidents, something bad seems to always be happening .

When is the last time you took a look at your disaster recovery planning? What do you expect would happen if the ceiling collapsed in your data center?

What would you do if an employee forgot to unplug a humidifier in his or her cubicle and the power grid feeding your servers and storage imploded?

This guide will cover five characteristics of data centers today that require new thinking when it comes to disaster recovery planning:1 . The new face of IT: Mixed platforms everywhere 2 . How virtualization can change everything 3 . Cloud computing and the new model of IT service

delivery4 . How to measure the return on investment (ROI) of

disaster recovery5 . Planning and testing for higher confidence

Soon, you’ll be ready to tackle these new realities of disaster recovery planning, with new approaches to protection that can make your life easier .

Don’t know? We are here to help.

Five Things You Need to Know Today About Disaster Recovery Planning

five things

4

Multiple Flavors Aren’t Just for Ice Cream AnymoreMulti-platform data centers are no longer an exception . They have become quite normal . Several years ago, most data centers were pretty homogeneous; all of the data residing there typically came from one vendor . This made things pretty easy to manage . There was only one set of instructions to learn and one phone number to call if anything went wrong . The problem was that this data also carried a massive price tag and was called a mainframe .

As the IT world continued to grow, several engineers and innovative companies (mostly out of California) started to design much cheaper processors, allowing other folks to build real servers out of PC components . These new servers used a new set of common standards, and they had a much lower cost barrier for entry, even for small organizations . Large organizations took advantage as well, and both had a buffet of building huge data centers with these x86 building blocks . Some call this the golden age of the data center . Analysts simply called it the client-server model .

The move from mainframes to smaller servers made the data center strategy simple: When you needed more applications, just bought more servers . Servers were cheap, so why not? The system worked well for a time, until the servers began to take up way too much space and use way too much power, even as most of the server resources weren’t even being used .

The solution to the problem was called virtualization . Virtualization allowed data centers to consolidate to a more manageable size and footprint . The problem was that all of these changes overlapped . Nothing ever went away completely . Data centers just got a bit smaller, or applications were shuffled around .

What is left today is a typical data center . It has been built up over the years with multiple flavors of platforms and vendors: a little bit of mainframe, a little bit of x86 and a whole lot of virtualization .

So what happens if the power goes out? Unfortunately, the hodge-podge of solutions in the data center means a little bit of this and a little bit of that for recovery planning as well . You would need one plan for the mainframe, one for the Linux servers, one for the Windows servers and probably a virtual recovery plan, too .

More flavors, more confusion, more headaches .

multi platform

5

Virtual This, Virtual That— Let’s Get RealHidden within the jumble of mixed-platform data centers is a secret . All those virtual resources can actually be used for more than simply virtual machine recovery . One wonderful thing about virtualization is that virtual machines can be used to run many different types of workloads . This flexibility can, and should, be applied to disaster recovery . Why not create a virtual recovery platform that can offer protection for all your workloads, including physical, virtual, Windows and Linux?

Virtual recovery plans can simplify, and in some cases eliminate, many of the platform-specific headaches of recovering from a disaster .

The typical process to recover a physical server is often: 1 . Find or buy equivalent or compatible physical

servers2 . Install an operating system 3 . Install patches and updates 4 . Install applicable applications 5 . Install patches and updates 6 . Upload backup data

Using virtual recovery, almost all those steps can be made a thing of the past . The process to recover a physical server can be as simple as: 1 . Power on virtual equivalent 2 . Be thankful you no longer have to deal with

physical servers

This is a far more elegant approach that can ultimately save a great deal of time as well .

virtuality

5

6

When Did Weather Metaphors Become So Common?Unless you’ve been living under a rock, you’ve probably heard that cloud computing is the next big thing . You might have even seen a TV advertisement or two talking about cloud computing . What is it, and can it affect, or better yet, improve disaster recovery?

We should probably begin with what cloud computing is . Some might say that when you look up a definition of cloud computing, you’ll find this: “buzzword .” But there are actually some well-accepted characteristics that define what is and is not considered to be a cloud of computers .

How I know my data center is ready to rain: 1 . Shared infrastructure (or, plays well with others) 2 . Rapid elasticity 3 . On-demand, self-service resource acquisition

model4 . Per-use billing 5 . Standard internet protocols 6 . Measurable

If you think about this, disaster recovery is a natural benefit of using this type of resource-consumption model . Disaster recovery is all about the unknown—not knowing when you might need extra resources, how many resources or for how long . The on-demand elastic model of cloud computing is perfect for those rather “cloudy” requirements, don’t you think?

But how do you get there? Most agree that the first stepping stone is virtualization . After all, virtualization is the main underlying technology that allows this cloud delivery model to be a reality .

So if you do have any aspirations for a cloudy disaster recovery future, the first step to take is to virtualize as many of the pieces that make up your disaster recovery plan and infrastructure today as possible .

thenextbigthing

77

Wait…This Sounds ExpensiveVirtualization has already proven its cost saving benefits in production data centers around the world . By reducing the sprawl created by years of accumulation of commodity x86 servers, virtual machines are now accepted as a cheaper way to run workloads . Not only do you dramatically lower your server hardware costs, but you also lower the cost of powering, cooling and maintaining them . However, despite these cost savings, there are still many barriers to fully adopting virtualization as the universal platform of most data centers . While everyone agrees that it is a cheaper way to run workloads, from a performance and security standpoint, there are still concerns around whether it is a better method .

Disaster recovery, on the other hand, faces far less scrutiny when it comes to infrastructure changes . As such, it is an area where a fully virtual strategy could be adopted immediately . Since disaster recovery is often forgotten, better performing infrastructure will be welcome by everyone involved, or sometimes the changes might even go unnoticed . (This is a good thing, by the way . Getting everyone to notice your brilliant new disaster recovery plan sometimes takes an actual disaster .)

But even if better performing recovery capabilities go unnoticed, the immediate cost savings won’t . As part of a disaster recovery plan, one virtual machine host could take the place of up to twenty or more

traditional standby physical servers . Think about it . The expensive duplicate infrastructure that data centers used to be need can now be a thing of the past . Eliminate the burden of keeping backup versions of each make, model and vendor version of all the servers you run . Replace all of it with a simple pool of virtual resources . This drastic reduction in the infrastructure has an immediate reduction in the cost of your disaster recovery plan .

But infrastructure isn’t the only thing that costs money . Your time is ultimately far more valuable than the gizmos you manage . Virtualization also reduces the overhead and labor associated with older backup and recovery-based protection approaches .

With virtual machines, everything from day-to-day maintenance, to recovery and even testing can be as simple as a few clicks of a mouse .

By implementing virtual disaster recovery, you can effectively get a double-whammy in terms of real costs savings . Cheaper to buy, and cheaper to run—what more could you ask?

costsavingsgalore

Why bother with convoluted, multi-step recovery processes

when you can streamline all your administrative tasks

virtually?

8

Hope Is Not a Strategy—the Hidden Secret of Disaster Recovery PlanningFor many years, the hidden secret of disaster recovery has been testing . Are you surprised? Think about it for a minute . When’s the last time you tested your ability to restore a workload? Now, how many did you test? And (here’s the kicker) what fraction of your total workloads was that? Backup and restoration is a real pain . Coordinating with everyone required from a facilities standpoint and getting everyone to agree on a schedule is hard enough, let alone having to rely on antiquated technologies like tape backup once you to get there . In cases where the difficulties associated with disaster recovery testing is too much, it simply never gets done .

The problem is that no testing or even infrequent testing means no confidence . After all, how can you be sure your plans will work if you don’t try them out once in a while? Even worse, in today’s business world, where everyone expects you to publish service levels and guarantee them no less, how can this be done without an accurate prediction of what a worst case scenario will look like? Users now demand accurate expectations . One or two days, or other ambiguous service levels are no longer acceptable . How can you publish an accurate number if you never run tests to prove it?

Wow, this DR stuff is really becoming a pain isn’t it? Again, the magic technology can be virtualization .

Virtualization eliminates the problems associated with a bare metal restore . There is no need to match hardware, or go through multiple steps just to get a test server up and running . This easy management allows you to simply select the virtual machines you want to test, create copies of them and power them on . Virtual networks also allow this process to be completely non-disruptive to your production processes .

Safe and easy—what more can you ask for? Easy testing seems like a simple concept . But the reality is, when you have real confidence in your DR plans, you will have confidence in setting service levels .

With the ability to easily and often run disaster recovery tests, you can accurately measure the time you expect to take recovering workloads when needed . Instead of hoping that you can recover from an outage in one or two days, now you can guarantee an accurate recovery time objective (RTO) and tell everyone about it, too!

This can even be used as a competitive advantage for your company . Everyone loves routine, so people want to know that in the event of a disaster, you can help them get back to routine as quickly as possible . Publishing service levels you can be confident in are tremendously valuable for any organization .

thehiddensecret

9

Fantastic! What’s Next?Go ahead and take some time to review your current disaster recovery plans . This new virtual world makes disaster recovery easy, but you still can’t move to it without some effort . Everything will take some time and planning . Use this guide as a bit of a jumpstart on what to be aware of, and where to start .

Try initiating some conversations within your organization . Ask people how important they are (everyone loves that) . This will help you build a better understanding of what the RTO needs really are for each workload running in the data center .

Use this simple uptime-downtime chart above to talk to your users about the dollar-cost impact these service levels might translate to . If your order processing application, for example, was classified as having three nines (0 .999) of importance, how many transactions do you expect would be lost or delayed, and what would that cost if you allowed for eight hours of downtime? Does that align with the cost of providing that three nines guarantee?

Remember, disaster recovery isn’t quite what it used to be . With mixed platforms, virtualization clouds, cost confusion and testing, there’s an awful lot to consider . Hopefully this guide has given you some food for thought as you consider updating or changing how your organization looks at disaster recovery .

Arm yourself with as much knowledge as you can find and start having some conversations about

disaster recovery in your organization . Now more than ever, it’s an important discussion to have, particularly given the number of changes we’ve talked about taking hold in data centers today .

Here are some additional disaster recovery resources from Novell:

Product information: http://www .novell .com/products/forge/ http://www .novell .com/products/protect/

Customer Success Stories: http://www .novell .com/success/hps_pharmacies .html http://www .novell .com/success/reed_smith .html

Whitepaper: www .novell .com/consolidated-dr-wp .pdf

Webinar: http://www .novell .com/media/dr

Allowable downtime for workloads:

Required Availability

90% (0 .9) 7889 36 .5 days

99% (0 .99) 8678 3 .6 days

99 .9% (0 .999) 8757 9 hours

99 .99% (0 .9999) 8765 50 minutes

99 .999% (0 .99999) 8766 5 minutes

Required Uptime hours/year Allowable Downtime /year

Arm yourself with as much knowledge as you can

find and start having some conversations about disaster recovery in your organization.