Amazon: Death by Cloud for Traditional Software


Enterprise computing is moving to the cloud, but few seem to appreciate just how profoundly this shift will destabilize the traditional vendor landscape. Gartner projects enterprise software to top $232 billion in global revenue in 2010, and it’s safe to assume that the Oracle (s orcl) and IBM (s ibm) crowd expect the lion’s share to land in their pockets.

Not so fast. Lost in the hype around cloud computing is the very real possibility that the company doing so much to enable it just might take a massive share: Amazon (s amzn).

This thought hit home for me while I was reading Redmonk analyst Stephen O’Grady’s excellent post on Amazon’s quiet revolution:

If Amazon restricted itself to basic public cloud computing services, that would be one thing. Most of the large systems players have turned their attention to the burgeoning market for quote unquote private cloud services. Whether these same cloud players appreciate the fact that a large portion of their interest in the private cloud is a function of the public cloud economic realities established by Amazon is unclear, but unimportant. Amazon is singularly responsible for the framing that is the public cloud today, a framing which generally relegates those with traditional enterprise margins in mind to private cloud settings.

But Amazon has not, of course, restricted itself to basic public cloud computing services. Amazon’s steady but underacknowledged expansion into adjacent markets is no secret. Since those early days when it took a Master’s degree to apply DNS to a running EC2 instance, Amazon has steadily grown its footprint beyond basic compute and storage functions into core enterprise software markets like messaging (SQS, SNS), analytics (Elastic MapReduce), monitoring (CloudWatch), and databases (SimpleDB, RDS). And, as of last Tuesday, enterprise Linux.

Amazon, in short, is an enterprise software company. It just doesn’t look or act like Oracle (s orcl), IBM (s ibm), or VMware (s vmw), so it gets treated as a side-show oddity when, in reality, it’s playing center stage. In many ways, it’s orchestrating the industry as a grand puppet-master, as O’Grady implies, and in the process, may end up disrupting established ways of doing business.

Amazon has started with infrastructure, virtualizing and delivering key components of enterprise data centers like databases, storage, and now Linux through its public cloud. In doing so, Amazon is not only shaking up outmoded delivery models like Microsoft’s (s masft) Windows, but also new approaches like Red Hat’s (s rht) Red Hat Enterprise Linux, as IBM’s Savio Rodrigues points out in his post over at InfoWorld.

Amazon is happy to effectively give away others’ businesses. As O’Grady illustrates, the entire reason we hear so much noise about private clouds from IBM, Oracle, VMware, Red Hat, and the like is that Amazon has effectively decimated their ability to make serious money in the public cloud.

So what happens when Amazon dismantles this wishful thinking and pushes all computing to its public cloud?

Such a move would naturally begin by Amazon getting enterprises comfortable running their applications on Amazon Web Services: companies like Pfizer (s pfe). Sure, today they insist they’d never run everything in the public cloud and, in fact, Eli Lilly (s lly) is an example of a company that has pulled back from AWS due to liability concerns. Clearly, it’s going to take time to get enterprises comfortable running their businesses on the public cloud.

But let’s get this straight: It’s just a matter of time. (s crm) proved enterprises will put their most sensitive data — their customer data — in the cloud. It’s unclear why a vast array of other applications and associated data would be any different.

Applications drive strategic value — and IT purchasing dollars — making it almost certain that Amazon’s support for infrastructure like Linux and MySQL is simply a way to enable the next round of its cloud strategy: application support. Over time, it’s hard to not see Amazon competing with the likes of and Google (s goog) in the cloud/SaaS-based delivery of enterprise applications.

So, today, Amazon is increasingly competing with the open-source vendors who sell support for a variety of open-source components of its Amazon Web Services offering. In the future, I expect we’ll see Amazon building out its AWS product portfolio in ways that make just about everyone in the traditional software market uncomfortable. Very uncomfortable.

Related content from GigaOM Pro (sub req’d):

Defining Internal Cloud Options: From Appistry to VMware
From Spots to Spikes: TV Taps Into the Cloud
Report: Evolution of the Private Cloud


James Colgan

Great to see this aspect of AWS’ strategy recognized. And you’re absolutely right – it’s just a matter of time. We (Xuropa) partnered with AWS and provide ISV’s and their customers an easy to use front-end so they can deliver their enterprise software without the need for any IT. So we’re in the thick of it. Discussing the transition to the public cloud with users of enterprise software, it’s already underway. We’re looking at a technology adoption curve just like we’ve seen over and over again – the chasm is yet to be crossed, but beach heads are well established.

Andy, CloudBerry Lab

I always enjoy learning what other people think about Amazon Web Services and how they use them. Check out my very own tool CloudBerry Explorer that helps manage S3 on Windows . It is freeware.


Amazon is a “retailer” the way that Berkshire Hathaway is an insurance company. Ie, it’s a core business and a major source of cash flow, but it doesn’t define what the company is.


Something to mention is that end users will still need to take care of the platform maintenance with Amazon EC2. The Windows Azure platform is taken care of by MS (patches, updates, load balancing, etc.) so the end user will just need to focus on the application. The relational database for Windows Azure is based on Windows SQL technology and provides high availability and Elastic Scaling with no need to understand usage or backup to scale out since it’s done automatically. Here’s a link with more details:

Jodi E.
Microsoft SMB Outreach Team


I apologize, my intention wasn’t to spam – I was trying to say that MS is also a viable option that lets developers just focus on their application without having to worry about the environment setup. Hope that clarifies.

Jodi E – MS SMB Outreach Team


This comment demonstrates Microsoft’s great presence in the cloud. Sometimes people get defensive when Microsoft is even mentioned. Jodi had valid points to make and included a reference to back it up. Failing to recognize Azure as a strong competitor to AWS and GAE in the cloud is a mistake.


Om, Why don’t we ever hear about Google App Engine on your blog. You seem to be slightly biased towards Amazon Web Services. Probably AWS is better than GAE for now, but what about the long term scenario?

Can we get your expert opinion on how this battle for cloud dominance will play out?


As a developer I have used both GAE and AWS. I feel there is almost no comparison. It’s like comparing general-purpose road automobiles (AWS) with high-speed trains (GAE). Use cases for the latter are always going to be far fewer.

If you happen to have an application that fits the rigid constraints imposed by GAE (Java or Python + Google’s MapReduce) then you have a real choice to make. Otherwise, and I think this applies to most Consumer and nearly all Enterprise Apps, GAE is never going to be a real option.
On the other hand, AWS offers the opportunity to do exactly what you are doing already but in a public or a private cloud. Which puts AWS in a different class.


Thanks Henchan. I agree with your analogy. Although, I have never used AWS, but have tinkered around with GAE.

But, I somehow feel Google understands the limitations and is slowly trying to add more service offerings like Google storage, Big Query etc. (although still limited to python/java). Plus Google Apps for enterprise is another way Google is trying to enter woo the enterprise market.

AWS has a huge lead and GAE still has quite a ways to go to even come close to matching the flexibility of AWS, but I wonder if they will jump the shark in the long run… kinda like they’ve almost done with android-vs-iphone battle.


Abstraction is one of the hardest things to get right. Also the most important. Google determined that programming language plus some custom data services was required. Amazon decided to abstract at the level of virtual servers. At first it was not clear which of these would win out but imho there is already a clear winner. Witness Oracle’s recent piling in to AWS and consider whether that would even be possible if AWS had GAE’s abstraction. GAE will not make up for this by adding services. In order to catch up with AWS it would need to be radically overhauled to support virtual servers.


Not just somewhat biased against Google – totally biased against Google ( may be it gets all the Apple eyeballs to the site). There’s no reason to not include Google App Engine in this article and if anything , I would also blame Google for not promoting or not showing commerical acumen in offering its services outside ads. Shame on Gigaom and Om Malik for being so openly biased against Google.

Bill McColl

Great article. The idea that many big IT vendors are putting out, that Amazon is a “retailer” and will quickly be overtaken, is a big mistake on their part. AWS is winning because it is exploiting three big factors – commodity hardware pricing, open source software, and startups. In the modern software world, where the game is “Developers, Developers, Developers”, this is a strong winning hand. Others can compete, but they will have to compete on all three fronts.

Most innovation in enterprise software today is coming from open source projects and from a new generation of enterprise software startups that are taking a “cloud-first” approach. In both cases, they really like what AWS offers, and what is offered just gets better every month.

At Cloudscale, we’ve built the world’s first Realtime Data Warehouse, and we’ve built it on AWS first. If you want it in-house to run on your own physical hardware, we can certainly provide that, but before you decide on that you should definitely check it out on standard AWS and on dedicated AWS clusters with 10GigE networking and high performance instances. These clusters are now even being used for high performance supercomputing in research labs, so they’re the real deal in terms of performance.

AWS has allowed us to quickly build a major next-gen enterprise software system that enables business users and IT teams to easily design and launch realtime, massively parallel big data analytics apps, and to do it all in the cloud. No hardware, no software, no hassle. Of course, if you really must “get physical” and go in-house (or “private cloud” as it’s euphemistically called) then you can, but over time (although not overnight) the reasons for doing that will gradually fade away.

Comments are closed.