Quick Thoughts: Oracle – When is the Cloud, Not a Cloud?
– Oracle has introduced new versions of its enterprise application software designed to be cloud-friendly, along with a revamped program for hosting services.
– Oracle’s “shared tenancy” approach is defensive, shielding its customer base from competitive inroads, but failing to deliver the benefits of a more aggressive “multi-tenancy” approach.
– Research on innovation (Christensen, Foster, et al.) has documented the grave risk to incumbents that take the defensive tack vs. emerging alternative paradigms. Oracle seems to fit the suit.
– While the new applications may sell well short term, bigger and more efficient operators will also compete to host Oracle’s own software, perhaps facilitating a future change to other platforms.
Oracle CEO Larry Ellison has swagger, along with his billions, his yacht racing career and the 20,000 twitter followers that he has amassed without tweeting. Last week, Ellison stopped in to the All Things Digital D10 conference to both scoff at cloud computing and to take credit for having invented it, all as a prelude to today’s official launch of Oracle’s fully cloud-enabled enterprise application suite and Larry’s first tweet. According to Ellison, Oracle’s approach to cloud software, which runs each customer’s software separately on its own virtual machine is superior to the software-as-a-service (SaaS) offerings of competitors like Salesforce.com in that it gives customers the flexibility to upgrade on their own schedule. Ellison claims that Oracle’s cloud move is the culmination of a 7 year effort to rewrite all of their applications to the requirements of the cloud, curious given that he spent most of those 7 years insisting that cloud computing wasn’t anything different than what Oracle was doing already. Of course, if cloud computing were just the shared tenant services and virtual hosting that Oracle is introducing today, Larry was right all along, as was arch-rival SAP, when it began promoting a similar approach more than 5 years ago.
Then again, there may be more to it than that. Oracle’s shared tenancy model essentially moves a client’s existing applications and infrastructure software – e.g. data base software, etc. – from their own data center onto a virtual machine running on one of Oracle’s data centers. The client can then downsize its own IT department and pay Oracle to manage and maintain the software. The alternative model is multi-tenancy – think Salesforce.com – which transitions the client’s data into a common software environment shared with other clients. Safeguards keep each client’s data secure from the others, but all of it is shared for maximum efficiency. Oracle pooh-poohs the multi-tenancy model as a matter of course, pointing out that it requires customers to change their software and lose the flexibility of managing upgrades on their own schedule, adding that their data is “intermingled” with the data of other customers to fan the flames for IT managers still unconvinced by the security measured implemented by cloud providers.
This is, of course, wildly self serving. The multi-tenant SaaS model, which Ellison so thoroughly distains, has significant advantages in scalability, cost and performance that can’t be washed away. For example, Salesforce.com can effectively provide CRM for enterprises with 5 employees up to many thousands. Because the software can be maintained centrally and support pooled, the cost can be dramatically lower, with the potential for more efficient utilization as well. SaaS also supports a more distributed data center architecture that reduces network error and latency, enabling superior performance for users. Moreover, SaaS can enable inter-organization collaboration in ways that are more difficult for virtually hosted applications. As security concerns over the cloud continue to recede, these advantages are a siren song to enterprises looking to minimize IT spending while improving application performance.
Unfortunately for Oracle, moving to a multi-tenant cloud application means moving data off of existing application and infrastructure software, opening the lucrative annuity of license fees and software support to the light of competition. Better to sow fear, uncertainty and doubt, keep customers tethered to their sunken Oracle investment, and gain a new revenue stream as a paid host for infrastructure and application software. This deflection is the classic incumbent’s response to innovation, documented ad infinitum in the work of Dr. Clay Christensen (The Innovator’s Dilemma) and Richard Foster (Innovation: The Attacker’s Advantage), amongst others. The problem is: defensive tactics may sustain an old architecture, but they also set the incumbent even farther behind once innovation achieves critical mass and begins driving the market transition in earnest.
Moreover, Oracle’s shared tenant solution may not be particularly competitive. Other cloud-based hosts, such as Amazon and Rackspace, are already supporting Oracle infrastructure and applications on a similar basis. Oracle’s rather cautious capital spending over the past few years compared to its hosting rivals suggests that its data centers may not be as state-of-the-art or as scale efficient as the company implies. The new software announced today is obviously an advance over previous Oracle ERP offerings, integrating web-friendly functions and designed for easy migration from private to public data center deployments, but it is not likely to be exclusive to Oracle’s own hosting operations, and given its shared-tenant design, may actually run better and more cheaply in someone else’s data center.
Finally, the biggest risk is that a customer that is convinced to take the plunge on public cloud hosting may be also induced to consider alternatives to Oracle’s data base software and applications. Exposed to the light of competition from far cheaper, open source based alternatives backed by support from a leading edge web host, some customers will likely bite the bullet and make the change. Customers that port their in-house Oracle implementations to 3rd party hosts, will hear the pitch to switch from their new partners as well. All of this weakens the traditional bond tying Oracle and its customers, and with it, Oracle’s ability to capture a pricing premium.
Of course, the threat to Oracle may be big, but it is also slow moving, and today’s announcements will likely sell well near term into a customer base that has a considerable sunk investment with Oracle and a propensity to take small steps rather than big leaps. However, with time, I expect enterprises will grow more comfortable with multi-tenant SaaS solutions, and one by one, customer defections will gain pace. While the timing on this transition is the critical question, Oracle’s dependence on the older paradigm and intransigence on the new one make it unlikely that it will be ready however quickly it plays out.
For our full research notes, please visit our published research site.