How Amazon Web Services Uses Formal Methods

2 weeks ago, Amazon appear its quarterly earnings, reporting a much larger net loss than expected. There was much speculation by pundits most the reasons for the scale of the loss (including me in a CNBC segment). Many commentators placed responsibility for size of the loss on Amazon Web Services – after AWS responded to an approximately xxx percent price cut by Google, the size of the "other" AWS category, in which Amazon places AWS revenues, savage iii percent from the previous quarter.

Subsequently the proclamation, a number of tech commentators discussed the implications of this 3 percent acquirement drop:

Quentin Hardy of The New York Times says that, despite the quarter "other" category being upward 35 percent over the previous twelvemonth, AWS growth may be slowing. At present that it faces additional competition from Google and Microsoft, along with the demand to do more than handholding of potential enterprise customers, it may face tough times alee.
Affront Darrow of GigaOM agrees that AWS faces competition from Google and Microsoft, simply notes that these companies are coming from style backside AWS; the potential for harm is prospective, not something that's happening today. Implied in her commentary is the fact that steep price cuts triggered by actions by Google and Microsoft take much more of an touch on AWS than on those companies exactly because their revenue base is so much smaller.
CRN UK notes that, with increased contest from Google and Microsoft, AWS can no longer maintain it has more deployed infrastructure than its nearest iv competitors combined.

Across these assessments, one well-known venture capitalist says AWS is overpriced. Beyond a certain amount of monthly spending, users are amend off building out their ain infrastructure. Moreover, he states, AWS offers terrible support and mediocre service offerings. Considering of Amazon'south disability to change its low-cost, low-touch nature, which he refers to as Amazon'south "Scorpion Problem," Google and Microsoft have a tremendous opportunity.

Co-ordinate to these commenters, Amazon faces a watershed moment. Earlier this quarter, effectively unopposed, it ran wild and built a huge concern; however, from at present on it faces unending competition, relentless pricing pressure that will atomic number 82 to dropping revenues and abandonment past customers seeking better quality solutions who also feel AWS is overpriced.

AWS Not Likely to Worry Near One Quarter of Revenue Drib

While all of these reactions to Amazon'south most recent quarter are, no incertitude, well considered, they're guilty of analyzing ane quarter's results and bold they predict serious long-term problem for AWS. Instead of foreseeing trouble on the horizon, y'all could have taken this quarter's Amazon results and interpreted them this way:

Taking a 30 percent price cutting and suffering only a 3 percent driblet in revenues implies, to my heed, around 25 percent quarter-over-quarter growth in usage. Annualized, that implies about a 100 per centum usage growth rate and, assuming that each quarter doesn't see a 30 percent price cut, an increase in revenues of around 40 percentage. This doesn't include any additional growth due to cost elasticity – of which, as this memo regarding its ebook pricing dispute with Hachette indicates, Amazon is keenly enlightened. Then the AWS usage growth rate might actually increase as a result of this price cut.
Don't count AWS out as a event of toll competition from its newly energized competitors Microsoft and Google. While they undoubtedly have deep pockets, Amazon historically focuses on driving prices lower and has enormous skills in increasing efficiencies and reducing cost factors. It's unlikely that it was surprised by a competitor using toll every bit a differentiator and has undoubtedly prepared for ongoing price competition.
The companies nigh damaged by this fierce price contest aren't the AWS, Google and Microsoft troika. It's everyone else. They view these low prices as an existential threat. (This piece, for example, bizarrely argues that depression cloud computing prices are a disservice to users, instead of more honestly saying that brutal price contest just exposes those unable to sustain their business absent high prices). You could view Rackspace'due south move to providing only managed cloud services as an admission that, like a Tour de France rider who falls back from a breakaway group to the peloton, information technology can no longer keep up with the tougher contest. The ongoing toll war has turned this market place into one that requires deep pockets, constant innovation, intense operational efficiency and steely competitive nerves. The biggest losers will be the would-be providers that find themselves in desperate financial condition as fourth dimension goes by.

Little noticed amid all the wailing (or gloating) about Amazon'due south AWS prospects is a report released recently by Pacific Crest Securities. Far from predicting doom for AWS, information technology states that AWS will exercise $v billion in revenues in 2014 and goes on to say, "AWS remains on a hypergrowth trajectory despite the law of large numbers and remains on pace to essentially double revenue every two years."

[ Case Study: How AWS Helped Scale App for World Cup Social Media Hub ]

This is, in effect, an upgrade from previous AWS revenue estimates. Last year, I gathered all the estimates I could find for AWS revenues, drawing on numbers from GigaOm, MorganStanley, MacQuairie and RW Baird. The blue line in the chart below shows those estimates, which placed potential AWS revenues in 2020 at $20 billion.

Amazon Web Services revenue projections to 2020 — AWS revenue, 2006 to 2020 (projected), in millions of dollars.

Pacific Crest'south numbers, identified past the red line, phone call for around 40 percent growth per year, in line with my previous signal. They come up out far larger – to around $forty billion, twice the previous estimates. This kind of increased opportunity for AWS (as noted by Gartner'south Lydia Leong and InformationWeek's Charlie Babcock), to my mind, proves more persuasive than the threnodies described earlier.

Of form, these are estimates. Amazon refuses to pause out bodily AWS revenues, preferring to bury them in its "other" revenue category. However, given AWS acquirement growth, we're probably only one or 2 years away from a time when financial regulations will strength Amazon to report actual AWS revenues. At that fourth dimension, all of this discussion volition move from speculation to fact, and nosotros'll know exactly how well AWS is doing – unlike some others in the space, who have had the SEC come up subsequently them about how they're accounting for what they dump into their cloud revenues.

The Real Message About AWS Revenues and Growth

Those who view this price state of war and acquirement guessing equally merely something confined to handicapping cloud provider competition make a serious fault. In my view, the biggest impact of the growth of the cloud provider market place is its effect on traditional Information technology and its vendors. R.Westward. Baird estimates that every $one spent on a deject provider displaces $four that would have been spent on traditional IT gear.

[ Analysis: Equally Cloud Apps Soar, CIOs Take on Deject Service Banker Function ]

Amazon's 2020 acquirement won't just be $40 billion spent on AWS. It will be $160 billion not spent with traditional vendors. As Leong notes, the ongoing growth and price reductions on the part of public cloud providers makes on-premises infrastructure – and therefore kit from legacy infrastructure vendors – decreasingly highly-seasoned. Cloud computing will dramatically change the nature of IT organizations as well, transforming their charter from nugget ownership to infrastructure management.

Then, while many viewed Amazon's Q2 fiscal results with glee and more than a little schadenfreude, information technology'southward besides early for AWS skeptics to break out the champagne. Instead, it'southward time to recognize where nosotros really sit down: On the crest of a sea change in the way IT is washed around the world.