October 2012

Business Continuity

Top Tips for Business Continuity

By: Stephen Jepson, Hardware.com

Ask yourself the question. Could my business survive after this?

During the Taiwanese floods of 2011 and despite one of its key sites being under 30ft of water, Western Digital remained in operation. It’s an often quoted statistic, but one worth repeating, that ‘around 60% of small businesses go out of business six months after a natural disaster such as a fire or flood’. Being a large multi-national, Western Digital have an inherent advantage in that they have data centres located around the world that can host data and applications that may have been hosted in Thailand. Whereas most small businesses don’t have such privileges they can take action to mitigate the outcome of any disaster. So here are Hardware.com’s five top tips ‘Business Continuity’ for the rest of us.

1. Be prepared
Plan for all eventualities large and small but allocate time and effort commensurate with the size of the problem caused by the disaster.

2. Be agile
Make sure the systems you deploy are capable of being used anywhere, using products like Vmware. Make sure you can use remote terminal or web sessions to access your critical applications.

3. Be realistic
Decide what level of data loss and what level of time outage your company can survive and trade it off against how much resource you want to put into a disaster recovery solution.

4. Be holistic
Business Continuity doesn’t depend on IT and data systems alone. If your people have nowhere to work in the event of a disaster then your IT DR planning may come to nothing.

5. Be prepared
I know I’ve already written about preparedness, but it’s the most important. Create your plan(s) and test them regularly. Be prepared to be flexible and deviate from any steps if it makes sense. Recovery Strategy Backup is only the first step in a Recovery Strategy. The ability to recover data within a particular time frame (Recovery Time Objective) from a particular point in time (Recovery Point Objective) is what any Recovery Strategy is about. Having said that, the first action in data recovery is to have at least one backup.

The second is to move it away from your main data store. In its simplest form this can be a removable disc or tape secured in the company secretary’s bag that is taken home at the weekend. In its more advanced form it is the replication of data over wide area links between data centres. The technologies involved here are:

  • The bag Despite the cost savings and simplicity of this method it is not the most secure and fastest way to recover systems.
  • The online backup Online backup services do offer a more advanced level of security, however these still involve slow restore times that depend on normal internet bandwidth.
  • Host replication Software such as Vmware Site Recovery Manager allow near real time mirroring of individual servers between data centres.
  • Storage replication Most entry level storage devices such as the HP P2000 G3 and P4000 include* inter-site replication to allow near real-time replication of entire data volumes over IP networks.

* Additional cost option with P2000. Included with P4000

The graph above shows the improvement in recovery times (RTO) against the cost of solution for various tiers of recovery strategy, from PTAM (pick-up truck access method) to automated, data replication and redundant data centres.

It could be argued that virtualization and its associated technologies are making the curve a lot shallower by bringing multi-site, near real-time recovery solutions within the range of SMEs.

Now, let’s examine data storage and how it can speed up data recovery times and mitigate loss in disaster situations.
Store more with less is the mantra on many storage vendors’ brochures but how do they achieve this?

Most vendors now embrace thin provisioning technology where the storage device only provides the blocks where data has been written while fooling the server that all its empty space is available as well.

Consider a 2TB data volume that has only 1TB of data written to it, which represents a 50% saving on the physical disc space. We have all experienced data compression utilities like WinZip on our desktops that compress files by eliminating empty space and recurring values within files, usually allowing us to email files that would otherwise fall foul of size limits. Some storage vendors such as NetApp and Tintri employ compression by caching and compressing the data in NVRAM before writing it to disc. The physical disc savings for compression depends on the susceptibility of the data to compression. Word documents will compress more than video files for example, but savings of 10% or more can be expected.

The same vendors save further space by de-duplicating data at the block level. Think of a file server with multiple copies of identical or similar documents. From the storage point of view, these are made up of many similar or identical blocks of data which can therefore be reduced to one block with multiple pointers so the other copies can be built from the one.

Hard discs are getting larger and cheaper so why store more with less, why not just store more with cheaper discs. By compressing and de-duplicating data we can keep more copies of our data and transport it over network links to more locations which in turn allows us to have much faster recovery times and recovery points to aid our business continuity plans. Also, no matter how fast discs reduce in price and increase in capacity we will still outstrip their capabilities by our desire to store more and larger data.

For more information regarding Business Continuity solutions, please visit: www.hardware.com.