Cloud Computing : Is it for us ?

I attended a very good session on cloud computing today. In the light of the Google POC that we might run this year, this was THE session to be in. Incidentally, the presenter, Bob Balaban, was from Binary Tree, a company which assists in migration of email, IM et al and to whom we have been talking about the Google POC!

After a high level introduction to cloud computing, we had an overview of the modus operandi of the major vendors "in this space"(*). Bob The Presenter was refreshingly non-Microsoft/IBM/Google biased but did drop numerous hints on what his company could do to help with migrations. We will be talking to them on vendor showcase day.

We also had a demo of a cool piece of code that migrated emails from Lotus Notes to Gmail - very satisfying for my inner geek - highlighting along the way this interesting fact : Google apparently throttles data input @ 1 transaction per second, which means that if we want to migrate 1,000,000 things (emails/pictures/attachments/... ), it will take us 278 days, ie 11.5 days!! In a public consumer context, this makes complete sense - around the world users upload single objects at a time so it makes sense to build the queue "horizontally" - go Democracy! In an enterprise migration, however, uploads will need to be done in batches for thousands of bursting inboxes. Defo needs some clarification from our Google friends when we meet them this Friday in New York, but in the meantime, can you please start cleaning your inboxes ;-)

The benefits of moving into the cloud are numerous and I'm sure you know them by heart - no kit inhouse, no cumbersome rich clients on users' machines if it's all browser-based, reduction in inhouse support headcount (ie errmmm the staff could be "re-deployed to perform other more value-added activities"), dead-easy scale-up or down on user base, etc. etc. The downsides are less numerous : migration costs have been known to outweigh any reductions in TCO, loss of control of the environment, co-existence costs and issues during the migration period, legal risks relating to the location of the data in the cloud deserves a whole blog by itself (http://www.gregonsecurity.mars/? ;-) and perhaps more importantly, quality of support service can (will!) suffer.

So - would Cloud Computing work for us? My initial feel is that it would, but only if we do it for our whole user base. Yes we might have a lower Quality of Service, Support & Availability, but would this not be insignificant enough to be an acceptable compromise given the savings per user involved? This is part of what we intend to find out more about with the Google POC. On the other hand, I think having part of our user population serviced on premise and the other part in the cloud will only drive costs up because we don't operate on a large enough scale to have a very flexibly scalable environment. We would very likely end up having a full-fledged premium environment inhouse anyway and pay extra for the lesser tiers of service!

Or is the question - if the world is heading to the clouds (and it definitely seems to be!), when is the right time for us to join the flow? So far there have been few large companies that have taken the leap, although a very well-known black-caffeine-filled-bubbly-liquid-that-comes-in-zero-or-full-fat-versions brewer recently moved onto MS Online. The question comes down to whether we want to be pioneers or followers, and if we want to be pioneers, can we afford to do so in the current economic climate?

The Google POC should hopefully allow us to answer some of these questions. Watch this space!


(*) As a side note, I'm pleased to report that I was not completely mad in my thinking in LoLA that IBM have left a gaping hole in BlueHouse - IBM's SaaS offering for online collaboration via web conferencing, social networking, IM, Shared Documents, ... - by not having any email functionality incorporated in it. They have now announced the acquisition of OutBlaze.com, whose messaging functionality will become an integral part of BH. I did mention this to anyone who would listen at LoLA, I'm sure the acquisition has nothing to do with what I said, but the whole thing does make me feel warmly smug ;-)