Why data protection must be central to any business continuity plan

By Chris Ross.

The Internet may have brought a lot of freedoms and a new way of doing business, but it has also ushered in a lot of new challenges. Perhaps the biggest challenge is the expectation, or even requirement, that a business is available to customers 24 hours a day, seven days a week. Lengthy downtime as a result of a disaster just is not feasible, which is why all companies need to implement a business continuity strategy to keep the organisation operating even under the most difficult of circumstances.

Central to this strategy should be a data protection strategy and backup regime. Everything that a business does is based on data and without this data a business will fail to keep operating. Protecting this data then, should be a primary concern even though it is also one of the most difficult jobs a company faces. If the increasing amounts of storage and smaller backup windows - due to the pressures of a business being available all of the time - were not enough, then there are additional pressures that have to be thought of, for example compliance with regulations such as Sarbanes-Oxley, Basel II and HIIPPA.. All businesses are responsible for how their data is stored and need to make sure that they stay legal when it comes to storing data, deleting it and even backing it up. It is clear then, that how storage and data is managed is vital to a company’s success. But it is important to remember how this data will be recovered in the event of an incident, rather than just how the data is going be backed up. In the face of a disaster, a business needs to get up and running quickly and efficiently; having its entire data sitting safely backed up but inaccessible is no good to anyone. Recovery times are as, if not more, important than backup times.

Business continuity versus disaster recovery
From a data storage point of view, in many ways business continuity is a similar concept to disaster recovery. Both strategies are concerned with protecting your data, but while disaster recovery is concerned with storing raw data and protecting it, a business continuity strategy should also take into account how a business operates. Business continuity should not just be concerned with raw data, but also its purpose and how it is linked to applications. The purpose of your strategy and the hardware and software you use to implement it is to restore data to a known state after an incident. When you restore data you should be restoring functionality and it is this concept that differentiates business continuity from traditional backup and recovery.

More than this, though, developing a business continuity strategy is not just about dealing with data as it stands on your network now, it is about getting your data in order before you start backing it up. Creating an ordered approach to your data before it is backed up is more efficient. By preparing your systems and choosing the data that you need to backup you can save time, money and disk space. This also makes it quicker to backup and, more importantly, quicker to restore, so you can react more quickly to an incident.

Ease of use
The most important part of managing your storage, whether it is on live systems or part of your backups, is ease of use. Storage requirements are growing at a phenomenal rate, so it is important that your management keeps up. As a result, scalability is also key and therefore you should try to understand how quickly your data centre will grow in terms of size, application coverage, etc. and look for solutions that can easily and non-disruptively grow with it. The ideal situation is to have one front-end that gives you complete control of your data. Disparate systems spread over multiple servers with multiple operating systems will not give you the view that you need, making it difficult to manage your network and increasing recovery times.

Choosing your backup software is crucial, then, if you are to make it easier to manage your environment. Planning for the future, not just now, is really important. For example, it may seem like a good idea now to buy a Windows-only solution for your servers. While this might be a good short-term investment, in the long term it is doomed to failure. IDC predicts that the Linux server market share based on unit sales will rise from 24 percent to 33 percent through 2007 alone. Saugatuck Technology predicts that nearly 50 percent of all enterprise data centres will use Linux to support mission critical applications by 2011. It stands to reason, then, that your company will be using Linux soon if it is not already. If you are locked into a backup solution that does not support heterogeneous environments then you are heading for trouble. When considering how to best manage your environment, it is essential that you plan for the future and understand that your environment is not going to consist of a single operating system.

Virtualization
The typical view of virtualization is server virtualization. It is one of those technologies that is constantly in the press and something that IT managers often get excited about. The concept is certainly a compelling one: a single server can be used to run many virtual servers, reducing management and the cost of hardware. In practice though, server virtualization is rarely pushed by business need, but tends to be implemented by IT Managers. This can be a dangerous situation to be in.
A virtual server requires pretty much the same amount of management as a real physical server; it still needs to be installed with the same software as a physical one, and this has to be licensed just like with a real server, and there is still the usual day-to-day administration. However server virtualization has the marked advantage of letting you get more from existing hardware; whereas without you would have required additional devices, by virtualizing servers you can maximise existing hardware and hence increase ROI and lower TCO. Licensing is probably the biggest headache as the current schemes are very confusing, particularly when a company offers discounts for running on a virtual machine.

However, this is a small price to pay when you think about the significant savings in space and particularly in power that will be realised over the long-term through having less hardware. The issue of power-saving is enticing, particularly with a push towards green computing.

From the point of view of business continuity, having virtual servers running off-site can save on the cost of setting up a disaster recovery site, as you need less hardware to run the same number of servers. However server virtualization needs to be considered carefully before it is implemented and certainly should not be used in the place of physical servers where you already have the hardware.

While server virtualization might not be ideal for all businesses, there is one virtualization technology that all enterprises should use: storage virtualization. It is a technology where you will get a real immediate benefit both in management, reduction in complexity and an increase in backup and recovery times. With storage virtualization you make your software independent from your hardware, making the most of your investment in both, as you can depreciate the cost of your software at a different rate to your hardware. The hardware independence is to be lauded for other reasons, too. Primarily, you are no longer locking yourself into one vendor that provides you with your backup solution. Instead, you can update your systems as needed.

The most common form of storage virtualization is Virtual Tape Libraries (VTL). These use hard disk arrays that ‘pretend’ they are tape libraries to backup software. The benefits are immediate and help reduce the burden of management. For starters, it is perfect for legacy hardware, as from its point of view it is still using tape, so there is no need to change your current backup regime. Newer systems can benefit, too, as your managers do not have to rethink their current backup strategy and can apply the same techniques they have used for tape to the new storage.
Even more important, as you have got a virtual layer between your backup device and your servers, you can change the backup device and the servers do not need to know anything about it as you are no longer tied into hardware. This could be for a bigger or faster hard disk array or, in the future, completely different technology. This hardware independence saves on what is perhaps one of the biggest headaches for network managers: data migration. In addition, it dramatically eases scalability and improves productivity because adding storage can happen non-disruptively without requiring downtime.

With tape backups, if the hardware is updated, then the old tapes have to be migrated to the new storage format. Not only is this difficult, but migration often introduces problems and raises the risk that you will lose data. When you are in the process of migration it can make it hard to recover data. Storage virtualization cuts out all of these problems, as you do not need to migrate data. Instead you can either keep your old system running, in which case it will be accessed in the same way as the new hardware thanks to the virtualization layer; or you can transfer the data to the new system through the virtualization layer. As both new and old hardware are communicating in the same way through this virtual layer, neither needs to understand how the other actually works or how the data has to be formatted.

A virtual tape library running on disk arrays is also a lot quicker than tape. For backups this means that the IT department can back up more data in less time, which is perfect considering that they are faced with ever shortening backup windows. Recovery is also a lot quicker too, so your business can be brought back much faster than if you were relying on tape.

It is important to remember that just because you are using hard disk arrays, it does not mean that you cannot archive important data to tape, though. You can still run tape drives off the back of your VTL for the information that you need to store long-term and off-site. As your virtualization layer is expecting to talk to a tape library, there is no complication in actually writing to tape rather than a VTL.

To implement virtual storage you are faced with two methods. The first is that your backup software includes the virtual layer. In many ways this is the easiest way to go, as you have then only got one piece of software to manage that can talk to your storage. The second method is to buy a third-party virtualization layer. This ties you to the virtualization software company, as you need the layer to talk to your backup media, but it means that you are not tied to hardware and you can still replace it when you need to.

While VTLs have traditionally been expensive products that only large enterprises have been able to afford, that is no longer the case. While they cannot be described as cheap, the cost has come down to the point where medium-sized enterprises can afford them. Expect prices to come down further in the next few years, too.

Replication
As important as backups are for business continuity and no matter how much better storage virtualization makes your data protection, you should be looking at replication, too. Replication takes an existing server and replicates the changes in real-time to another server. What you get, then, are two servers with exactly the same state every time any change is made. One of these servers can be in your production environment and another one can even be off-site, across continents, ready and waiting for when it is needed, such as your current office being unusable.

The advantages of replication are huge when compared to backups. For starters, there is no recovery time involved. The second you detect an incident you can fail-over to another server, which is already ready to take over with a consistent set of data. With backups the best you will get, even with a regular backup policy, is data that is a couple of hours old.

Backups also require the data to be restored and it may not be in the exact format that your applications are expecting. Replication fixes this by replicating data in the exact format that an application is expecting. As one of the key drivers for business continuity is to think about the applications and how they are tied to data, replication clearly solves this problem. It is a much more flexible solution for critical systems and can give you a situation where you have got zero downtime.

In general, then, replication works well alongside backups. You use replication for the critical systems, while backups give you a structured and layered copy of all data and applications in your business. Together they allow your business to respond quickly and efficiently to any incident.

Be clever with your data
Storage virtualization and replication both provide good ways of protecting your company and speeding up recovery in the event of an incident, but they are only as good as the data that you store. Before you start backing up and replicating everything you need to start thinking about what is actually being stored and being smart with what and how you backup. Storage resource management (SRM) is a way of doing this.

An SRM system gives you a console that lets you see exactly what is being stored on your network. It provides a consolidated view of all your servers and backup devices, which you cannot do easily without SRM; it is not easy for an administrator to look at 30 servers and provide an accurate account of what is being stored, where it is located and how many duplicates you have got. By knowing exactly which files you have got and where they are stored you can intelligently manage you backups to make best use of available resources.

For example, you might not want to make backups of all the MP3 files stored on your network. As well as being a waste of space, there is also the amount of time it takes to back up and recover this data in the event of a problem. By managing what gets backed up and when, you can make more efficient use of your backup window. As you are also using your backup media more efficiently, too, making better use of the investment you have put into it, you may even be able to prolong the life of your existing systems, putting off expanding your storage until you need it for business use.

Cutting out unnecessary file types is not all that SRM is good at. The enterprise-wide view it gives you allows you to see all of the copies of a file on your network. Rather than make six copies of one file, for example, SRM allows you to make just a single copy. Again, you are cutting down on backup and recovery times and saving on storage. All the while you are making it easier for your administrators to view storage across the whole network.

There has been talk of how information lifecycle management (ILM) will help drive SRM, but this is not technically true. SRM is something that all companies should be interested in, as it makes business continuity easier and more efficient; ILM is really a different subject and its requirements will really depend on your type of business.

ILM is concerned with data protection, how long an item of data can be stored for and when it has to be deleted. This is particularly important with new regulations that make it a legal requirement to store some records and another requirement to delete other records. Failing to do so, even on a long-forgotten tape in a warehouse is illegal. SRM in itself will not make this job any easier, as it is not interested in the contents of a file, just that the file exists, where it is stored and how many copies of it exist. SRM looks at the outside of a file, while ILM is concerned with the data that is inside.

Ultimately, it may be possible for SRM systems to look inside files and provide ILM as well, but we are probably many years away from that scenario. In the meantime, if you need ILM you will have to buy software that enforces your policies independently. It may be that you can operate both systems side-by-side, though: if your ILM system knows which files have to be deleted or backed up, your SRM system can immediately report where those files are and do the job for you. Just remember, though, that SRM is concerned with managing your storage more efficiently, while ILM is concerned with managing the information stored. Buy each system with that in mind and do not expect an all-in-one solution. Do invest in SRM, though, as being able to track all of your storage from a single console makes management easier and, ultimately, recovery times quicker.

How to choose your data business continuity plan
Talking about what needs to be done and actually doing it are two different things and the job can be complicated even when you have identified your needs. For starters, there is the problem of all of the different systems that you will have in your data centre. It is a simple fact that no one supplier will have a single solution for your particular business continuity strategy. The right products need to be chosen for the right environments; you might be best looking at a solution that covers 80 percent of your company, as this will give you the easiest management; the other 20 percent will have to be treated differently.

Here, though, the problems often set in and it is difficult for any IT department to get it right, not because they do not know what they are doing, but because they will only ever have to come up with one business continuity strategy. The new technologies, such as replication, are harder to implement than traditional backup as it is not your IT department’s day-in-day-out job, picking the right solution is really difficult and it is something that you do not want to get wrong.

For complex systems like this the channel has a lot to offer. The right channel partner will have more experience in business continuity strategies than your company and can advise and consult on the right solution for you. They will be able to implement the many systems that you need and choose the right products so that you can manage it all in the simplest way possible. The expertise that you will find in the channel will give your business the best chance of bouncing back after an incident quickly.

It is common to worry about how a channel partner will be able to adapt its skills for different types of businesses, but the mechanics of it are not that difficult. Every company has two requirements when it comes to business continuity: how long can the company afford to be down for and how much data can it afford to lose? Some companies cannot tolerate any data loss, but they can afford to be down for a couple of days if that is what it takes. Other companies cannot afford to be down at all, but can stand to lose 20 percent of their data.

No matter which market you are in, the slider will always be set between these two points. As long as your channel partner knows where to set that slider and which data is critical, they will be able to come up with the solution for you. Planning this and working out your requirements will ensure that you get the business continuity strategy that you need.

Conclusion
Business continuity is not just something that big business should be thinking about; it is something that should be central to all businesses. Key to its success is your data protection scheme. How you implement this can mean the difference between a company’s success and its failure. We have seen, though, that it is not just about being able to recover from a disaster, it is recovering quickly and efficiently that counts. Key to this is tying your information to your applications; it is not just about storing raw data.

When you are planning your business continuity strategy, it is important not to get too carried away by buzz-word technologies. Only implement a technology if it provides an improvement over your old systems, not just because it sounds interesting. Technologies such as storage virtualization can help you achieve your goals, as they make backup and recovery times quicker, and give your software independence from your hardware. It makes perfect business sense, too, as you can depreciate your software at a different rate to your hardware.

Managing your storage intelligently is crucial too. SRM ensures that you are only backing up data that needs to backed up and you are not duplicating files. It makes the most efficient and cost-effective use of the resources you have and makes your storage easier to manage. Your ultimate aim is to be able to manage all of your systems through one console; the ease of use this gives makes business continuity easier to implement and your company quicker to restore. With the right channel partner and your company’s specific set of requirements, a business continuity strategy will protect your data and make sure you continue operating under all circumstances.

Chris Ross is VP and GM, BakBone Software EMEA www.bakbone.com

•Date: 16th October 2007• Region: World •Type: Article •Topic: IT continuity
Rate this article or make a comment - click here