Overcoming Virtual Challenges

Over the last few weeks, I've been doing an informal survey of IT organizations aggressively pursuing virtualization, as part of some other work I'm doing.

I wanted to share with you the common themes that I'm finding over and over again -- you either know these things already, or -- perhaps -- might find them interesting in your particular situation.

Background

We wanted to know how people are overcoming specific challenges.  We went to EMC's "Studio E" group who manages references for us, and asked them to send out a short survey.  BTW, if you're an EMC customer, and haven't joined Studio E, you might want to consider it ...

Lots of responses.  I winnowed out about a dozen I found interesting for one reason or another, and got them on the phone for about 30 minutes each just so I could get a bit more color.

The questions were as follows:

#1 -- How does your organization view virtualization?  A great way to save money, a better way to deliver IT services, or a bit of both?

#2 -- Much is being made of the need for new operational models in fully virtualized environments.  Do you agree, and how might your organization be addresing this need?

#3 -- Part of any cloud model -- internal or external -- is shifting to a "pay for what you use" model.  Where is your thinking on that topic, and how do you see your organization tackling that issue?

#4 -- As many people have found out, there can be strong resistance to going to a fully virtualized environment.  What have your experiences been, and how are you overcoming resistance?

#5 -- Part of the pitch for private clouds is that IT organizations will have the flexibility towork with compatible service providers using compatible technology, yet under the control of IT.  What are your thoughts on this?

#6 -- VMware, Cisco and EMC recently made a big announcement that they were working together to align roadmpas, build new solutions (i.e. Vblock) and create joint services offerings.  What are the pros and cons of this -- from your perspective?

#7 -- Going forward, do you see technology issues being the limiting factor in IT evolution, or will organizational issues be more important?

#8 -- Any other thoughts you'd like to share?

Compared to other surveys you might see, this one is very anecdotal and has no supporting methodology to speak of.  The customers -- while all EMC customers -- were based in North America, and covered a variety of industries and company sizes.

My real goal?  Find people who had interesting perspectives on the above questions, and talk to them.  Maybe about these issues, maybe about other ones -- it didn't matter to me.

Make of this what you will.

It All Started When ...

In most of the interviews, it comes out that a new CIO came on recently, and that's when things got kicked into high gear.  Sometimes the new person was an IT professional ; more often, they came from the business side.

This matches with my own perceptions in working with customers.  Lots of new leadership in IT organizations these days.  This means that the business, generally speaking, isn't happy with what they're getting from the existing IT leadership, and feels a change is in order.

Conclusion: this kind of changes the traditional answer to "what does the CIO want?".  If they're the new person in the role, they usually want to make big changes in a big hurry.

Org Challenges Vary By Size

One of the questions above speaks directly to the observation that -- in a fully virtualized environment -- processes tend to change, and -- as a result -- roles and responsibilities tend to change as well.

Although everyone violently agreed with the underlying premise around "how you do IT changes in a virtual environment", there was strong division as to the seriousness of this effect.

Smaller, more nimble IT organizations just adjusted a few things, and got on with it.  Larger organizations, however, found themselves hamstrung by existing org models and skill sets -- not to mention inherent resistance to change as well as people defending their turf.

Conclusion: smaller IT organizations seem to have a big advantage in fully exploiting virtualization to run IT differently.

Resistance Is Futile

Very few of the interviewees spoke of organizational resistance to using virtual servers vs. physical ones.  Now, of course, there were always a handful of applications that didn't make sense to virtualize, but -- with everyone I talked to -- that sort of server-hugging had largely subsided.

The real resistance?  Software vendors who weren't comfortable with running their product on VMware.  That being said, most interviewees saw it as more of a nuisance than a problem.  In addition to support FUD, everyone agreed that licensing schemes were still firmly rooted in the physical world, and that was a problem.

Conclusion: software vendors not being on board -- either support-wise or licensing-wise -- is a major pain in the patootie, but isn't really slowing things down, based on the people I spoke to.  They wish we could get past this and move forward, though.  And no one is blaming VMware for this.

Be Ready For Proliferation

Nearly everyone spoke of the explosive proliferation that can result with server virtualization.  And it makes sense -- when servers are easier to create, people create more of them.

If you think about it, physical server resources are shared.  Storage and software licenses aren't -- and that where things could get out of hand.

Some organizations hadn't put any effective governance in place, and now were struggling to tame the new virtual beast.  Others had created lightweight "approval committees" for all new requests that were, in reality, nothing more than quick email dialogs.  And a few had simple chargedback for any licenses and/or storage consumed for virtual servers.

Conclusion: if you haven't seen a proliferation problem already, you will, so be ready for it.

Making It Easy For The Business

One overwhelming theme I heard in every interview was the need to make things as easy as possible for the business.  They all understood that the majority of the appeal in external clouds and services wasn't that they were cheap -- it's that they were easy.

Just about everyone had serious initiatives to take the friction out of "what the business wants" and "what the business gets".  It ain't all about cost reduction with these people -- it's about better, more responsive IT.

Conclusion: is cloud more about making things easy, rather than making things cheap?

So, How Do You Feel About (External) Clouds?

No one was in a hurry to use external service providers, with one notable exception discussed later.

If I were to attempt to paraphrase a general theme, it'd go something like this: we're not going to add external providers to the mix until we get our own house in order.  Put differently: once we feel we can control our internal IT to our satisfaction, we'd be glad to start considering external service providers.

This is a very different sentiment than the previous outsourcing sentiment from several years back, e.g. "your mess for less".  Willingness to use external service providers was all predicated around having control and visibility: not only into their own environment, but the service provider's as well.

That being said, I presented an imaginary scenario where service providers were doing a few things they're not doing today, and re-phrased the question.

The answer was quite different: everyone had a list of things they'd consider handing to an external service provider -- once those prerequisites were in place.  Conversely, there was also a list of things that would be the very *last* things to be handed over.

Conclusion: external service providers really haven't figured out what enterprise IT organizations want.  But it's a pretty short list, IMHO.

Chargeback is Situational

As we got into various chargeback discussions, another theme emerged -- it's nothing more than a tool to help govern IT consumption -- and little more.  Business users ought to know what things cost, and chargeback was simply one way of achieving that.

No one was really interested in granular, highly accurate chargeback models with intricate policies -- just something that could grossly estimate who was using what, and what the pieces might cost.

Conclusion: chargeback is another tool in the toolbox, and not an end in itself.  Unless you're a service provider, that is!

You're Doing What?

For every rule, there's an exception.  I had the pleasure of talking to a decently-sized IT organization that was using Google Apps for productivity and collaboration: word processing, spreadsheet, forms, lightweight workflow, collaboration, etc.

When I used the term "private cloud", that's what they thought of.  The knowledge worker environment was functional, secure and robust enough to meet their needs.  Not only that, the users loved it!

Now, these folks weren't in a highly-regulated environment, nor were they working with unduly sensitive information.  But it did make me take notice -- these people had bailed on the Microsoft and IBM (Lotus Notes) world, and hadn't regretted a moment of it.

VCE Reactions

Now, here's where sampling bias kicks in.  Remember, I was talking to customers of both EMC and VMware, so -- predictably -- they were relatively enthused about VCE, Vblocks, etc. -- as long as the combination created new combinations, and didn't take anything away from what they already had.

A few really understood what it was all about, and the potential.  Others were simply curious.  Just about everyone realized that Cisco was bringing something new and differentiated to the table with UCS, and they'd have to go learn more about it.

No surprises here, from my perspective.

Final Thoughts

I got to talk to some pretty smart and passionate IT people who were in the trenches.  They were justifiably proud of the progress they'd made, but realized there was much more to do.

And it was a pleasure and privilege speaking with each of you ... thanks!

Comments

[32gb ssd]

I am writing a research paper on virtual team challenges and this is one of the things which I am going to add in my paper. I really appreciate that you discussed this problem over a blog and will look forward to here some more of your experiences on virtual teams.