Last week Amazon Web Services brought its Enterprise Summit to London. The agenda promised customer stories by the bucketload, and we weren’t disappointed. For an event so keenly focused on persuading enterprise customers to stop worrying and learn to love the cloud, AWS sensibly let its customers do a lot of the talking for them.
We were, of course, regaled with roadmaps, reasons, and ways to transition to cloud from Amazon’s execs, but it was the customers who brought it to life with stories of Sunday mornings reclaimed, Christmases saved, and data scientists with time on their hands.
News UK, we heard, found (perhaps unsurprisingly) that its 150,000 Sunday Times subscribers all wanted to log in and get their content between 8 o’clock and 10 o’clock every Sunday morning. And with the company’s existing systems creaking at 17 transactions a second that meant no lie-ins for the IT team, as they spent every Sunday morning on-call to nurse things along when needed. But rather than ask the CFO for $1m to furnish a new data centre (as might have been the case in times of old), self-confessed "risk averse, suit-wearing, enterprise guy" IT Director Chris Birch recounted how instead the company plumped for the cloud. With AWS’ help, News UK were up and running in only 8 weeks and, from those beginnings in 2010, Amazon now provides over 35% of the company’s computing (and autoscaling has given the IT team their Sunday mornings back as they no longer need to worry about that peak load!). Cost-wise, Birch is taking a “10 year view for adoption” but is already seeing a gap open up between his spending on AWS now and the escalating costs of what he would have had to shell out, had he continued to tread his traditional (risk averse, suit wearing…) enterprise data centre path.
Online retailer, Shop Direct, revealed that its front end availability was painfully grazing 57.47% during December 2012′s peak buying period. Fast-forward twelve months, though, to the same time last year (after the company jumped cloudwards with AWS), and now it’s up to four 9s—giving CIO Andy Wolfe a much less stressful run-up to Christmas.
And whilst all these stories of CIO relief at improved performance and availability, and CFO-friendly cost savings, were interesting (and no doubt designed to instil an “if everyone else is doing it, why not me” feeling in the assembled enterprise masses), I confess I was really there to pick up what I could on any Big Data successes…. and here the Financial Times’ CTO John O’Donovan came through with some great numbers. The company adopted AWS Redshift for its data warehousing and saw dramatic performance improvements—its 40 data scientists apparently left “wondering what to do with their time” as their reports ran up to 40x faster in the cloud (and worked out 80% cheaper too).
So clearly there are gains to be had, if you manage your transition right. And on that point I’ll give the last words to AWS’ Global Director of Professional Services, Todd Weatherby. He was talking about how successes like those outlined above risk encouraging under-managed adoption across an enterprise, as access to a cloud platform is rolled out and everyone wants a piece of the capabilities, the capacity, and the cost reductions. He cited examples of where previously tightly-buttoned enterprise IT gets relaxed a little too much, too quickly—allowing all sorts of inconsistencies to creep in. His answer: despite temptations to over-do the whole ‘new way of working’ bit, it still pays to “be a bit prescriptive” from the off, because “nobody wants to circle back” and start again properly once there’s a head of steam built up.
Agile’s good; anarchy, not so much.