With its entrance into infrastructure-as-a-service Oracle is not attacking Amazon Web Services, rather it is erecting a defensive wall to stop existing customers leaving Oracle for the AWS cloud.
With Ellison's announcement on Sunday that the database giant was going to launch an infrastructure-as-a-service component of its Oracle cloud, many media publications saw this as a sign that Oracle was going head to head with Amazon in an attempt to gain market share in the cloud.
This is the wrong way of looking at it. After talking with analysts, Oracle executives and reading between the lines of statements made in keynotes at the Oracle OpenWorld show, I have come to believe that Oracle's infrastructure cloud endeavour is nothing more than an attempt by the company to defend itself from losing its existing customers to Amazon.
Oracle's cloud is for existing customers, not non-Oracle developers
Oracle's announcement today of a cloud storage service and plans for an equivalent compute service by the start of 2013 is not an attack on Amazon's low-cost cloud built on commodity hardware. Instead it is a tool Oracle can use to to convince existing customers to not shift to a hybrid cloud format where they use Oracle software and hardware internally, but go to Amazon to run their database workloads.
"We have a lot of customers in the world who run our databases and middleware... the first goal of the Oracle cloud is to appeal to those customers," Oracle's executive vice president of product development, Thomas Kurian, said when I asked him if he could give me a reason why a non-Oracle developer using the Amazon S3 storage service would want to move to Oracle's storage service.
Competition with Amazon may come later, but for now it seems Oracle is building a cathedral for its faithful, and steering clear of entering the brutal price war entailed by competing with the 'slum' clouds operated by Google, Microsoft, and Amazon.
Hardware dependence could make it difficult to compete with Amazon on price
"Do we aim to be price competitive with the market at large? Our view of why we are getting into IaaS is really more to do with what we are hearing from customers - that they really need a full stack offering," Abhay Parasnis, senior vice president of Oracle Cloud, said when I asked him about whether Oracle would be competitive in terms of pricing with Google, Microsoft and Amazon.
Though Parasnis acknowledged that "price and economics is a key imperative" of IaaS, he did not give any pricing information on the service. When pressed, he said Oracle will eventually "be price competitive in the market".
I am not sure how this is achievable; Oracle's dependence on its expensive engineered systems could make it very difficult for the company to compete with Amazon, Google and Microsoft on price.
"If by using larger systems they can reduce even further the cost (and hardware is only a fraction of the total cost - power, cooling, building, people, networks etc) and initiate a price war in this space then they have a shot if they go at scale," Simon Wardley, a researcher for the Leading Edge Forum, told me via e-mail. "However, everyone else in the industry focuses on reasonably sized highly commoditised systems because this is where they've found the best price points, and Oracle's competitors have lots of experience in doing this."
Google and Amazon are known to design non-standard hardware and complicated secret software systems to bring the cost of their clouds down. Oracle, on the other hand, is committed to running its cloud on exactly the same infrastructure that it sells to customers, and there is no evidence yet that this can compete with commodity gear and secret software on price.
A cloud to defend the database
It seems to me that the IaaS component of the Oracle Public Cloud is nothing more than a defensive move. For this reason, I think it is wrong to assert that Oracle is now in competition with Amazon - it is not in competition, it is rather building a cloud for existing customers purely to stop them migrating Amazon. If you are not an Oracle customer, there is nothing, yet, that is welcoming about Oracle's IaaS proposition.
Evan Quinn, a senior analyst for the Enterprise Strategy Group believes Oracle is doing IaaS to 'shore up' its database technology and retain customers.
"Competition among the IT superpowers has to do with what I call "control points," for example Google leads in web search, a major control point; Apple and Google lead in smartphones because they open the primary operating system control points; Oracle has maintained its leadership position for decades because its database rates as a major enterprise IT control point," Quinn says.
"Possessing a control point gives the owning vendor the power position in a large ecosystem – all other vendors must somehow pass through the owner, or ignore the owner at its own considerable risk."
For this reason, Oracle is building out its cloud to help it defend its "primary database control point," Quinn says. Again, the IaaS announcements are defensive, rather than aggressive moves, it seems.
A dark cloud on the horizon
My take is that Oracle's lacklustre approach to IaaS could ultimately backfire. Unless the company releases convincing pricing and performance figures, or a unique technology, it is unlikely developers will be swayed away from Amazon.
Though Oracle stands to make money off of existing customers with its cloud I don't believe this is a tenable long term strategy. Unless Oracle puts a lot more effort into the development of the IaaS component, then I think it's safe to assume that the cloud is a band-aid to staunch the flow of customers into Amazon and nothing more.