This article is more than 1 year old
Attack! Run. WTF? A decade of enterprise class fear and uncertainty with AWS
How 2006 was like 1914 for Microsoft, IBM, Oracle et al
Ten years ago, Oracle was mid-snack, taking a break between swallowing PeopleSoft for $10bn and Sun Microsystems for $8.5bn.
Microsoft had settled a long-running feud with Sun over something petty involving Java for $20m and had begun designing its vision of how the internet should be programmed with its new BFF, IBM.
Microsoft and IBM were crafting the Web Services family of XML-based specs with Plus One Other Vendor [insert name here] they deigned to invite to their late-night study group.
IBM was the cool enterprise IT kid Microsoft wanted to be, so hanging out over the WS specs was cool, right?
Like the imperial powers of Europe of centuries gone by colonising Africa, these great powers and others in tech were carving up and "reimagining" the enterprise and the web the way they felt it should be.
The only unknown was LAMP – Linux, Apache Web Servers, MySQL database and Perl/PHP or Python – but even that wasn't a big threat.
This tiny uprising proved containable as its leaders were soon co-opted – made to run on either their platforms or their middleware.
The great powers didn’t know it, but they were about to face their own First World War event: something was about to strike their world that would destroy complacency, reset reality and leave the old men of tech gasping for breath.
It was Amazon Web Services, marking its 10th anniversary this week.
Today, AWS operates in 190 countries, and has thousands of customers – in business, that is, its online services span the gamut – having begun in 2006 with a teeny, tiny 5GB Simple Storage Service (S3).
In those 10 years, AWS has rolled beyond S3 to include compute, data warehousing, content delivery and storage.
But AWS has not been confined to merely boring infrastructure, it has been expanding into applications like analytics and it has hosted the software of those elder statesmen - Microsoft’s SQL Server, Oracle’s database and ERP software from SAP.
And if open-source was running anywhere, it was running on AWS. So, too, were and are all those sexy frameworks that succeeded LAMP as well as a good number of app startups.
Pay as You Go
Jeff Bezos' web services unit delivered not just a technology platform but way of consuming business tech: paying for what you consume using clearly published units of payment.
Amazon has now driven down the price it charges for those units a whopping 51 times.
Customers in the UK today include the Co-op Bank, News UK, Financial Times, British Gas, National Rail Enquiries, National Trust, Netflix, Shop Direct and many more.
In the US, among many, many more, you'll find the CIA – snatched from IBM, which was so flabbergasted that the Blue Shirts of Armonk decided on the old-school route to victory and filed a legal complaint asking the government to re-evaluate IBM's deal against that of Amazon, which Big Blue later withdrew.
Gartner now reckons AWS is the number-one provider of infrastructure as a service. Translated, this means the Amazon wing has become a staple provider of enterprise technology. It is doing this in a market once defined by Microsoft, Oracle, IBM and Sun – to name just four.
The Sun sets on Oracle threat to AWS
Sun Microsystems, of course, no longer exists. The sad irony is Sun was the first to actually talk about offering the kind of utility computing that AWS now offers.
Oracle, which bought Sun, preferred to play a Game of Thrones that was corporate M&A to hold onto its position in IT.
Sadly, it chose wrong; Oracle spent $8.5bn on Sun but ultimately discontinued the company’s fledgling utility computing service.
Hardware and Java was what Oracle wanted.
Today, Oracle's resultant hardware business makes just half the revenue of AWS and is is shrinking - falling 13 per cent to $1.1bn - versus AWS's 69 per cent growth last quarter to $2.4bn.
That past complacency of Oracle’s CEO on cloud has put Oracle firmly in a pack of also rans behind AWS on platform cloud, with Oracle now throwing PR at a problem to convince Wall St it is credible as a provider of IT as a service.
Microsoft was remarkably quick in responding to AWS compared to its technology peers – unveiling its Windows cloud, Azure, just two years after AWS. But, being Microsoft, Azure succumbed to classic innovator's dilemma: how to sell a new platform as a package and at a price to maximize revenue without cannibalising the company’s actual main money-makers – PC and server software.
After delayed starts under Ray Ozzie and Bob Muglia, the technology roadmap only really clicked under new CEO Satya Nadella and executive software nerd Scott Guthrie. One brought the CEO-level commitment, the other made Azure work for developers.
Gartner today regards Azure as number two, behind AWS, and yet... According to Gartner’s incumbent Cloud Queen Lydia Leong, Azure lacks the polish of AWS.
Microsoft is significantly behind [AWS] but is closing the gap,” Leong told The Reg last year. “Microsoft will trade on being ‘good enough’.”
IBM got caught looking in the other direction, too. Famed for re-inventing itself around software in the 1990s under Lou Gerstner, the majority of IBM’s focus for the 2000s was devoted to unloading the PC and the server businesses on China. The firm is now trapped in a maelstrom of transition, restructuring and layoffs.
Like Microsoft, IBM seems to have believed AWS couldn’t happen to it, that what the world needed was the same server software and services.
It was nearly seven years after AWS that IBM realised something was afoot – probably when it lost both the CIA deal and got slapped about its attempts to make the CIA love it – that Big Blue said it would spend $2bn buying computing player SoftLayer and in 2014 throw $1.2bn into a massive data centre expansion to host your data and compute.
It’s very easy to see how the imperial gentlemen of IT blundered into this WWI.
AWS, after all, is the offspring of Amazon, a company that started in 1995 selling books – books! In other words, a consumer dot-com. But if its boss Jeff Bezos proved anything, it was he had ambition and a simple philosophy to keep trying – even if things didn’t work and even if trying meant hurting Amazon’s profit.
And whereas old-school bookseller Barnes & Noble stuck to selling books in its belated response to Amazon, Bezos was soon off into other fields – within two years of going online, Amazon was selling online music and by 1999 toys, electronics, tools and hardware. Amazon went public in 1997 and was the first internet retailer with one million customers. For a long time there was one e-commerce game in town and – barring eBay and Martha Lane Fox’s Last Minute – it was about the only dot-com to survive the terrible millennial implosion.
Amazon went though its own pain, losses and restructuring, late and badly fulfilled orders during the early 2000s, but the firm has obviously endured.
The movement into new markets finally brought Amazon to the point where we pick up this story, when the firm began selling excess capacity on the one thing it had lots of - servers - in 2006. That was storage-on-demand and, two years later, compute - EC2.
The momentum of all such surprise attacks can never persist. The great powers suffered their WWI, but the enemy’s troops are now out in the open.
What next? History dictates that those attacked try to consolidate their positions and respond.
AWS is still attacking – growing at a phenomenal rate, 71 per cent in its recent quarter to $2.4bn and 69 per cent for the year to $7.88bn. The appetite among enterprises for AWS’s style of technology and model of delivery clearly hasn’t yet been satiated.
But that means there’s opportunity for those competitors, who are now forming ranks.
Microsoft is number two – according to Gartner – but, hey, who needs perfection when you’ve got brand, channel and enterprise relationships to leverage - as is the case for Microsoft? Let AWS be Betamax on technology, Microsoft can be VHS - even if it can’t beat AWS on price.
Oracle also has the brand and the accounts plus software-as-a-service offerings that are growing as fast as those from rivals. Also, it has the cash to survive at least a few painful quarters or years of transition - going so far this week to declare stock dividends and a $10bn buyback.
The only question is how long Oracle's Wall St overseers are willing to put with a management unable to make its core business grow.
The problem for Oracle rapidly growing that platform business that lags Amazon and that Amazon has now extended into Oracle's core of databases.
Also, retaining cloud sign-ups and turning them into long-term paying customers – an area where The Reg understands Oracle is struggling. Dealing with that is a structural problem of how Oracle sells cloud in the first place - never mind the actual quality of the technology of its cloud. If Microsoft lags AWS in terms of polish and finish, then where exactly is Oracle?
Of these, IBM looks weakest: January saw a 15th consecutive quarter of falling revenue and income with the cash-cow software business down six per cent. There are massive layoffs and a management captivated by the media reflection of its PR fairy of Artificial Intelligence.
Cloud is still nascent at IBM. Cloud is to IBM what “other” was on the Amazon reporting sheet before AWS was broken out as a separate line item last year.
Cloud is part of an “other” category at IBM, something the firm’s management calls “strategic initiatives”, which also includes analytics, mobile, social and security. As a whole, this lot accounted for $29bn, up 26 per cent, or 35 per cent of overall revenue.
For all this, however, Amazon seems to recognize a challenge must come.
Migrate to see you
It’ll soon become easier to shovel more data into AWS, as Amazon chose the 10th anniversary of AWS to take its bulk upload AWS Database Migration Service out of preview and to make it generally available. Finally, a bigger shovel for the customer.
AWS has another aspect on its side: the more data shovelled in and the more services it releases, the more entwined customers become, and the harder it becomes for them to declare UDI.
Rare, indeed, is the example of Dropbox, which moved all its data to OpenStack, a shift people seem to see as a sign of hope. The average enterprise won’t have the time or budget to simply move its data. It’ll be perceived as about as valuable to the company’s strategic growth as moving off of Windows XP. Which is to say, something given a very low priority.
Like it or not, even if the old guard snapped back tomorrow, the truth is AWS now has its fences across so much of the cloud, removing them isn’t an option.
The big question then for AWS at the age of 10 is this: when will the old men of IT regain their wind? How big will be their counter-attack and will it be concerted? Will it pose a tangible threat and how would AWS respond?
Right now, the opposition seems more concerned at taking out each other: Microsoft pinching Linux and Oracle’s on-prem database users for its own SQL Server rather than throwing what it's got at AWS. Yet, it's AWS's Relational Database as a service that's now Amazon's fastest-growing service - taken over from Redshift. Will the opposition's lack of appetite for a head-on fight change?
Of greater concern for AWS is getting new customers stay. Yes, there are plenty of big names in the AWS bed but we hear, too, of those who test AWS just for its cloudiness or for dev and test and then retreating to private cloud.
The big question, then: will it be sufficient over the next 10 years for AWS to simply offer yet more new technology and yet more price cuts if it wants to keep the fences rolling out in the faces of the old guard? ®