Virtualisation will be a key infrastructure priority for chief information officers looking to reduce costs and increase energy efficiency over the next 12 months.
Three-quarters of ZDNet UK's sister site silicon.com's 12-strong CIO Jury IT user panel said virtualisation — the simultaneous running of several operating systems on a single desktop or server — is one of their current investment priorities.
Peter Pedersen, chief technology officer at Rank Group, said: "Virtualisation is very much a priority to reduce cost, to increase agility, reduce space requirement and to reduce 'carbon footprint' through reduction in power utilisation for both equipment and cooling."
Reducing the number of physical servers is also saving on management and facility costs for Ben Booth, global chief technology officer at polling and market research group Ipsos.
It's good for business, as we can cut costs, and good for the environment, as overall it helps us reduce our carbon footprint
Ben Booth, global chief technology officer, Ipsos
He said: "It's good for business, as we can cut costs, and good for the environment, as overall it helps us reduce our carbon footprint. Of course, major applications still require multiple servers but, for the smaller systems, it is proving to be a very effective technology."
Cost savings and energy efficiency aren't the only benefits of the virtualisation approach. Paul Hopkins, IT director at Newcastle University, said: "There are lots of different and often complementary reasons, including capital cost savings, revenue savings, green issues, resilience, performance, and capacity management."
Graham Yellowley, director of technology services at investment bank Mitsubishi UFJ Securities International, added: "We are looking to leverage virtualisation to maximise our computing efficiency, reduce our data centre space, power and cooling requirements, and to provide an effective disaster recovery facility at a cost lower than purchasing multiple individual machines. Combining multiple grid computing farms and adding virtualisation is also on the agenda."
One company already well down the virtualisation path is insurance group Markel International. Steve Fountain, IT director at Markel International, said: "We actually embarked on this some three years ago and have already 'virtualised' a very high proportion of our Windows servers onto a very much smaller number of 'significant' IBM Intel servers. Very successful it has been too, so much so that VMware is now our stated 'preferred' platform of choice."
But not everyone is convinced by the arguments for virtualisation. Mark Beattie, head of IT at LondonWaste, said: "We have so many different systems which need to interact, I just don't have the confidence that it'd all hang together if we tried virtualisation."
Talkback
It is so refreshing to see a technological emergence coinciding with a cultural shift under the umbrella of 'Green IT'!
With the public, private and governance pressures arising from realisation of our environmental responsibilities we are finally witnessing the erosion of the cultural inertia which stopped the economic and functional benefits of such innovation being recognised.
So - if people can see the benefits from using virtualisation tools and approaches for consolidation (yes - I think that really is all we are talking about here!), does anyone think we are ready to finally wake up to the fact that we do not actually need to have a physical desktop at every desk? ... or, heaven forbid, that we can access our logical desktops remotely from practically anywhere? …
… just think about the energy savings that would "reduced costs and increased energy efficiency" that would give!
Let's see. Virtualisation and consolidation. The new total solution hypes.
From my own experience, the only environments that actually improve are badly management and/or designed ones. But those are easy to find nowadays.
cost reduction:
- Knowing the initial investment required to even get started, and knowing the investments needed to stay up-to-date, not by a long shot when compared to well designed and managed environments.
What's the ROI? One year, two years, longer? Don't you think that by then something new comes along that starts a cascade effect?
increased agility:
- Ever heard of OTAP? The kind of acceptance test that should be done when making changes in highly interconnected and complex environments are so time consuming that most resort to fire and forget deployments. Meaning, if nobody screams hard enough and can't be silenced enough it's the new standard. Issues are creatively "solved" on a need to basis with hind sight. One size fits all is the mantra, all bow. Again, well designed and managed environments don't have such problems.
reduced space environment:
- Meaning? By putting all eggs in one basket, concentrating them to the max and making them highly interconnected and depending to each other the risks increase ten fold. Previous small disasters become huge disasters, the need for back-up, fall-back, redundancy (servers, networks, power, climate, etc), etc solutions increases. So where's the reduced space environment? Again...
reduced 'carbon footprint'
- Don't think so. Often lots of hardware that could have been used for years to come is thrown away because it's not part of the new standardization. That's a lot of garbage that needs to be taken care of and new equipment doesn't exactly grow of trees. Again...
reduction in management and facility cost
- Usually the opposite. The need to acquire new tools to keep control and insight increase. For one because (small) mistakes are no longer easily forgiven by the environment (another drawback of over centralized total solutions). Again...
other benefits
- Basically any overhaul should lead to performance improvement and other benefits at first. Question is, what happens as time passes? Usually the situation slowly degrades until some critical threshold is reached and collapse sets in, followed by panic and symptom fighting. After that the situation rarely returns to what is once was. Again...
In all, well designed and managed systems take in account the needs of the many AND the needs of many ones. With damage and change durability designed into the environment right from the beginning and guarded thereafter. Straight jacket solutions that are the dream of one department but another's nightmare are not a total solution.
Loosely connected decentralized environments have much more, damage and change, durability with fall-back and reduced state capabilities, are more forgiving and give better value for money, certainly in the longer run, when well managed and designed centrally.
KISS means so much more then thinking with just one hat on. It's putting all the pieces together with future proof in mind. Seeing the entire picture, over time, as a, e.g., field full of domino blocks. Thinking about, what would happen if a push this domino block? Or this one? Or those three together? How to know? How to stop in time? How to recover? What if I throw in a box of new domino blocks? Etc. Much easier to accomplish once you start thinking outside of the box. Doable if you can simplify across the board.
Until then, hope that the commercially motivated give you what you really need. History shows that they won't. They're much more interested in: will you pay me now, or will you pay me later?
Also, we need to get rid of previous century concepts. Who cares about a desktop? Information (not data) and services from any place, anywhere with any device sounds much more future proof to me.
One size fits all solutions can never hope to acco
This post has been removed by a moderator.
A growing number of businesses are adopting server virtualisation to reduce costs and increase efficiency. In fact, IDC predicts that by 2011, over 1.1 million virtualised servers will have shipped. While adoption rates are set to soar, it is important that businesses put the right processes and tools in place to ensure the successful deployment of virtualised servers connected to SANs.
Adoption of virtual server environments is somewhat inhibited by the limited flexibility with respect to storage access. All virtual machines on a physical server currently share the same external storage resources and a single set of FC addresses provided by the underlying OS or hypervisor. This limits storage management, QoS and resource accounting capabilities and gives rise to data protection concerns in the SAN.