Original URL: https://www.theregister.com/2009/05/22/it_infrastructure_evolution/

When rip and replace is not an option

Unification doesn't have to mean blind conformity

By Jon Collins

Posted in Channel, 22nd May 2009 10:02 GMT

Comment Today’s brochures about data centre design show pictures of row upon row of rack-mounted devices, and come with the clear message that in networking, servers and storage, conformity is the answer, efficiency is the goal and value is the inevitable result.

This message is reinforced by many parties. Networking companies such as Cisco parade unified computing infrastructure strategy (I use the term ‘networking company’ guardedly, given that its recent announcements encompass servers and storage as well as networking.) Meanwhile, virtualisation vendors such as VMware argue that the brave new world of IT will be entirely virtualised.

Unification, consolidation, simplification – it sounds attractive. But not when you take the real world into account.

For a start, most data centres are a work in progress, tending to evolve rather than respond to any wholesale change. This is not an argument against consolidation, merely an observation that big-ticket, green-field deals are less likely than deployments that add to the ongoing evolution of the data centre - especially in the current financial climate.

Evolution, then, not revolution. This may sound like a broken record but it is consistently what we uncover in our community research activities. Here is a case in point: in The Register’s Barometer study of server hardware, “continued committed” use was indicated across a wide range of non-commodity (aka “legacy”) boxes. We see similar findings in all domains of technology. Whatever ‘stuff’ is already in place, little will be thrown away to make way for the next few pallet loads.

All the same, organisations want to take advantage of new technologies, to move information around and crunch numbers at higher speeds. When rip-and-replace is not an option, anything that is brought in must work with whatever’s already in place, and furthermore, it needs to make what exists, work better than before.

This puts a different slant on what can reasonably be meant by unification, in the context of IT. If the perfect-world view is impossible, then questions become more about how to make old work with new, and how to ensure that whatever is bought can have as long a lifetime as possible.

We absolutely see a desire to standardise, to simplify, to make more efficient, but only in the context of when this makes sense for the business. This gives us the best place to start: as we summarise in our paper “Taking Stock of the IT environment”, an understanding of what people need is pre-requisite to delivering cost effective IT. From this point, it is possible to assess what’s already in place, and look for justifiable improvements that can integrate appropriately with what is already there.

Overall, unified IT is as much about taking a unified view, as it is about delivering unified technology. There are many, very interesting developments in IT at the moment, from virtualisation to cloud computing. But in no case do they mean throwing out the baby with the bathwater.

Take Red Hat’s stance on virtualisation, for example, which allows both OS-based and Hypervisor-based virtual machines to co-exist in the same server pool. As for Cloud, it is unlikely in the extreme that any large organisation will adopt it wholesale, a factor reflected in the quite sensible positions of vendors such as IBM, Microsoft and (while it is still around) Sun.

The action point here is: don’t panic. Yes, absolutely, look at what opportunities there are to save money and deliver greater value through IT. But don’t be fooled that anyone has yet come up with an IT vision that really can replace everything that came before it. ®