Time to start regarding them as legacy?
If you dig back through historical media coverage from around 2007/08 you’ll find a lot of mentions of how x86 virtualisation was changing the world. The name of the game was server consolidation, and IT teams were waking up to both the cost saving benefits and the prospect of eliminating a lot of server admin drudgery. VMware was on the rise and Microsoft was just entering the fray, while Citrix and others were jostling for position.
Over the following few years, server virtualisation moved from promising new idea to accepted mainstream technology, and IT organisations got on with cleaning up their x86 server sprawl.
A lot of the systems put in back then still exist in one form or another, as do the practices that grew up around them. Even where upgrades and growth have taken place, the consolidation mindset with which many old VMware and Hyper-V environments were conceived still defines their spirit and the way they operate. The ‘discrete-fixed-stack-per-application’ philosophy is alive and well in these systems.
None of this matters that much if you are supporting a stable set of legacy applications with predictable resource needs. Granted you might get irritated by some software vendors stinging you for excessive licence fees as they continue to penalise customers who virtualise in many scenarios, but on the whole, you’re generally not held back in any significant way.
According to a recent survey, however, problems arise when you attempt to build on these past investments to create an environment suitable for running more dynamic workloads. We’re talking here about unpredictable, fast-growing or (potentially) hyper-scale applications, and/or those requiring secure distributed access.
Cloud infrastructure and platform services from the likes of Amazon, Microsoft and Google come into their own here, but you don’t always want to run stuff on other peoples’ servers – survey respondents were very vocal on this. Many, therefore, highlighted a need for cloud-style application delivery but with systems deployed on their own infrastructure.
The question is then whether you should extend the scope of the abovementioned older virtualisation environments, or start afresh with a platform specifically designed to create private clouds or support hybrid deployment models. The consensus is that in most cases, starting afresh is the way to go for a couple of main reasons.
Firstly, the existing technology stack you have in place is likely to need add-ons and extensions in order to deal with the management and orchestration needs of more dynamic applications, and these can be expensive. Secondly, getting the most from modern private/hybrid cloud platforms requires a mindset shift and the adoption of new ways of working, both of which can be easier if you aren’t fighting against existing systems and processes that were designed around more static requirements.
As a simple example, if your organisation is moving more towards Agile, DevOps and Continuous Delivery, things like self-service provisioning of resources by development and application teams, while maintaining operational integrity, is a key requirement. So too is the ability to move workloads around with ease, e.g. as demands change over the lifetime of a modern ‘digital’ application.
Against this background, we couldn’t help contemplating that it’s time to start thinking about all of those 10-year-old virtualisation environments as legacy.
The overriding conclusion from the research is that there’s a lot to consider when it comes to implementing on-prem platforms capable of supporting today’s more dynamic application and workload needs, but established vendors often put their interests ahead of yours when providing guidance and solutions. It can be better for incumbent vendors to encourage you to build on existing investments because selling add-ons is good for (their) business, and it keeps you ‘in the fold’ as a loyal or captive customer. The last thing they want to propose is a fresh start, as this might prompt you to then look around at other options.
If you are interested in some insights and tips on how to navigate this minefield, then you’ll find lots to help in our research report entitled “The Economics of Application Platforms” which is available for download here.