A business continuity/disaster recovery group on LinkedIn posted an interesting question about whether business continuity/disaster recovery was purely an insurance issue – or whether there is real return on investment.
We have been involved in BC and DR technologies that provide what is called “fault tolerant” computing for several years. These not only address the technical aspects of continuous processing, no matter what the circumstances, but also have a compelling ROI angle also. Hopefully the next paragraph or two will explain why…
Fault tolerant computing means literally that – should components of a multi server infrastructure have a fault, or indeed should a full host server have a complete outage – there is no impact at all on the computing environments – ie zero downtime. Even multiple faults can occur on separate servers in a protected pool at the same time, and again there is no impact on the processing environment. (example – two servers working together in our simplest of architectures – we can tolerate losing the complete disk array on server one, at the same time as losing the network connection on server two – and there is no impact at all on the computing environments that rely on this two server setup, not a single ping is dropped, no data is lost, users do not know any faults have occurred). This whole locally protected environment can then be replicated real time off site to complete the end-to-end local (BC) and off site (DR) protection. Most importantly from the pure technical BC and DR perspective, local technical outages such as disk failure or loss of network connection are treated as such and are protected – and only full site outages (ie a true site disaster) invoke the DR failover to a second site (hopefully never ! – so I accept the DR aspect of what I am describing is more of an insurance with less ROI argument). This technology also involves server virtualisation techniques – hence environments that traditionally live in a physical world – ie one operating system / computing environment per physical server, can be concatenated and operate on a much smaller hardware (and hence carbon and energy spend) footprint.
So what you may say – sounds like VMware or similar right? why get involved in the discussion and where is the point ref ROI etc? Well no it isn’t – it is much more cost effective and easy to use than VMware, and does not rely on expensive shared storage. This is on standard hardware with no shared storage.
So the question of getting senior management to buy into a BC / DR strategy – when there is no apparent ROI but perception is purely that BC / DR is an insurance….. Certainly, in the technologies we implement for clients – this is not the case. Critical environments such as security access, prison CCTV systems, banking and trading hall systems, shop floor manufacturing systems, on line ticketing / booking systems, casino systems, etc stay up and running with absolute zero downtime – (ie the primary objective). However, server farms previously of (eg) 50 physical servers can be consolidated down to just 6 or 8 host machines. It doesn’t take a genius to work out there are huge savings in terms of energy spend, air conditioning requirements, cost of lifecycle replacements, plus reduced and simpler maintenance and support overheads to look after such a reduced and simpler infrastructure. This is NOT cloud computing by the way – this is within the client’s own server room or data centre environment.
So at least as far as our customers are concerned – ROI comes into the discussion very early on, and tends to get senior management very interested indeed. In fact – it’s pretty much a 50 / 50 split with the entry point of project sponsorship between the IT manager/director responsible and the FD or MD. There are attractive benefits for all.