The changing expectations for modern backup & Disaster Recovery

In the aftermath there will inevitably be much recrimination, discussion and debate as to what constitutes the most effective DR strategy to deal with a catastrophic event.

Thankfully, in the UK, we’re unlikely to experience a natural disaster on such a large scale. You are more likely to lose access to your building because the guy driving the digger outside your window is unwittingly about to rupture a water or gas main, or rip through your electricity mains cable or communication line. Or the Air Conditioning in your server room goes horribly.

Regardless – these interruptions can extend to days, even weeks and will have a serious impact on your business and your ability to service your customers. In our intensely competitive commercial world, few organisations can afford to be out of business for a few hours, let alone three or four days. But how do we motivate non-IT managers and Directors to take the lead in assessing the risks and defining the plans and solutions to protect their organisations? The traditional view that IT staff should shoulder all the responsibility for business continuity should be consigned to the trash bin. Sadly, in the aftermath of any disruption, it’s invariably the IT guys who get the blame.

A new generation of ‘risk managers’ and auditors are asking uncomfortable questions about the resilience of organisations to withstand a prolonged disruption; more specifically – what are the financial and reputational consequences for each day that the organisation is dysfunctional? How exactly would they recover their IT infrastructure; and how long would it take to become fully operational? Hence there is very strong evidence from the marketplace that expectations are changing rapidly.

For those with traditional DR contracts, they can at least give an answer as to when they will be able to recover. Although if they are still using Tape based systems – it may not exactly be the answer people want to hear.

Traditional disaster recovery ensured that on average it would take anything from 24 hours to three days to get your data back on to new servers and that it would be done from tape. It was also based on the provision of shared hardware supplied within hours of a disaster, the assumption that the tapes didn’t have any errors and that the recovery process was straight forward where everything ‘restored’ without any issues.

But we live in a ‘Now’ society where IT systems are regarded as a utility and expected to function 24/7. Today if you say “it will take us 3 - 4 hours to recover”, the majority of business directors would simply not accept that, such is the changing level of tolerance or intolerance in our intensely competitive world.

Everybody has their own collection of DR horror stories and ‘4am in the morning’ anecdotes where the tapes were corrupt, or they couldn’t recover to new hardware. Hence, very few people will ever declare any love for Tape based recovery systems. The arrival of a new breed of Disk-to-Disk backup technologies was welcomed about four or five years ago and achieved considerable success on the simple concept that recovery at LAN Speed from Disk was eminently faster and more reliable than tape, which indeed it was - and it set a new standard in recovery time expectation of 6-8 hours. Those vendors offering BMR (Bare Metal Restore) within their technology could realistically claim to reduce this time by 30%. Problem solved ?? Not for everybody.

The more technically informed business users say; “We have Cluster technology within the production server environment – so why can’t we have remote Fail-Over systems for DR, similar to the banks and financial institutions?”.........Why not......”because it’s too expensive for an organisation of this size” was the response from most IT Managers.

The impression of “Too expensive for small companies” was historically true; if one explored the costs of High-Availability backup & replication software where vendors did not allow for any variation in license cost for critical V’s non critical servers. It was an expensive ‘one-price-fits-all’ licensing model. These vendors figured that if instant recovery was a business imperative, the customers would be willing to pay a higher premium.
But nothing lasts for long in our industry where young companies arrive with new ideas and innovative technologies and break the monopoly of the major players. These new ideas were based on logical extensions to the concepts of Virtualised architectures. So why not Virtualised Backup & Recovery with the ability to create a dedicated Virtual Clone of each production server that maintains an ‘always-on-ready-to-run’ status. But crucially, providing support for both physical & virtual servers and treating each server as a physical. Why is this so important: logical or not, most IT Managers will still maintain a number of critical servers as stand-alone physicals.

These basic concepts have been around for a while, but the final products have been elegantly perfected from the original idea of SAN to SAN replication and booting to a VM Host. Why is it selling? – performance and price. For the same cost as a conventional Disk-to-Disk system, it delivers on the simple expectation that Critical Servers should be fully operational in five minutes or less, day or night. For business users, it delivers on their ‘Now’ expectation that instant Fail-Over should be a basic prerequisite in any modern IT infrastructure.

Surprisingly, the costs are broadly equivalent to the standard Disk-to-Disk vendors, and less expensive when supplied as an appliance with storage and archiving. Great attention has been paid to the need for remote management capability. The elegance and simplicity of the Web based management consoles is quite impressive; designed such that any junior in the IT department could literally handle any recovery or DR crisis. The notion that you can get an alert on your Smartphone while you sitting on a train or out to dinner – and resolve it in 2 minutes is sheer heaven to any IT Manager who previously had to find his group of recovery experts and get them all to assemble on site before anything could be recovered.

As these systems develop and mature, they will create a highly competitive marketplace that will force the established backup vendors to completely re-engineer their now rather ancient architectures.