Oracle Cloud: 5 Big Challenges

Oracle has successful open source products within its offering, such as MySQL, and is a noted contributor to the Linux Kernel Development Process through the offices of Wim Coekaerts.

It also has as its legacy the driving of the Apache Harmony programmers out of the Java Community and its iron-fisted control over Java itself. MySQL has four forked projects, an unusual number, including Web Scale SQL, backed by Facebook, Google, LinkedIn, and Twitter. Amazon is offering another fork, MariaDB, backed by Michael Widenius and other original MySQL founders, due to distrust of Oracle's MySQL stewardship.

Challenge 3. Cloud Hands Power to the Customer

Amazon keeps artfully thinking of new ways IT users can self-provision a new service in the cloud and tie it to their legacy systems. It makes it simple and cheap to do so, with recent examples being its Machine Learning Services and Data Pipe streaming service. Oracle has the technical chops to do the same, as long as it's playing catchup, but does it have the culture to invent new, user-friendly services that can be self-provisioned at the pace Amazon is demonstrating?

Oracle, since the departure of the tough but civilized Ray Lane as president in 2000, has reverted to the kind of brash and confident company that doesn't hesitate to tell customers what they're doing wrong, to conduct surprise audits and present customers with a large bill at the end of the year, and to issue "take it or leave it" ultimatums when its consultants get into a dispute with project managers. Hand off power to the customer? To some extent the Oracle culture, more than other successful businesses, has denied power to the customer and taken their money anyway. In the cloud era, it's time for a personality transplant, but that operation is not yet available.

Challenge 4. Cloud Is Commodity Hardware

Oracle is counting on a dual hardware cloud, and only IBM has tried this, mixing x86 with Power servers for its large Power customer base. Oracle predicts the cloud market will be a consumer of its Sparc chips. It has invested in a breakthrough, highly secure design for its M7 chip. Executive VP John Fowler said the cloud represents a new opportunity for Oracle to find a larger market for Sparc, not only in servers sold for enterprises on-premises, but also in the thousands that will be needed in Oracle Cloud data centers. This indeed would be a new market, provided Oracle customers wish to convert to Solaris or keep running Solaris in the cloud as well as on-premises.

Oracle must convince them to do so at a time when the x86 cloud is proving more secure than many critics expected. A well-managed, highly uniform, and controlled x86 environment appears to be secure enough for many of today's cloud users.

Oracle faces an uphill fight in the conversion battle. It must convince enough customers to adopt Sparc, based on the new security features, before Intel finds a way to duplicate or match them in some future generation of Xeon chips, which would render the Oracle advantage moot. Superior Oracle hardware -- and right now, the M7 is a dazzlingly capable chip -- must be something other than the lowest-cost commodity hardware on the market. But so far the cloud has been a ruthlessly commodity construct. Solaris already is available in the cloud. Just ask Joyent. But it hasn't set the world afire. Can Oracle Solaris plus M7? We'll see.

Challenge 5. Cloud Is Credible

The whole relationship between the cloud provider and consumer is based on a high degree of trust. It is an agreement that invokes connections between remote data centers, credit card accounts, workloads that must self-provision, and two parties that must work on technical support far apart from each other. The respective data center managers may never even meet. Amazon has mastered this relationship, and Google and Microsoft show every sign of doing so. But if you're expecting the kind of cloud I've described, will you get it from Oracle? Only Oracle customers can answer that question.

Consider another question: Would you have bought Linux from former Microsoft chief Steve Ballmer? Ballmer dissed Linux and open source as "a cancer" and insisted Linux was violating Microsoft patents. If Ballmer were still at Microsoft, would customers be sending their Linux workloads to run on Azure the way they're starting to under Satya Nadella? Somehow, it seems to me that the Oracle Cloud is going to require more of a change than Larry Ellison changing his mind.