The Empire Strikes Back... almost...
Earlier this week, during the Welcome Keynote by Oracle CEO Larry Ellison at the Oracle OpenWorld, Oracle introduced the world to Oracle Exalogic Elastic Cloud. As a side note, the Infosys CEO was there as well at the event. The OEEC is a complete system of servers, network, storage, VM, operating system and middleware, all engineered to work together thus providing the foundation for a highly elastic, super-fast, secure and proven platform to host the entire enterprise application suite. The proposition being offered is simple. With cost containment being a major driver for most IT organizations, there is an increasing trend towards opting for setting up a private Cloud. It is this market that Oracle is trying to target by offering a machine that can host all the applications that would usually reside on a multiple hosts and allow the administering and metering capabilities that are crucial for operating the virtualized eco-system. The OEEC is a package of hardware and software components boasting of a 64-bit x86 processors, an InfiniBand-based I/O fabric and solid-state storage with Oracle WebLogic Server, other enterprise Java Oracle middleware and a choice of Solaris or Linux. Due to this, it offers the computing capacity required for hosting a private Cloud.
My view
Oracle has long been trying to convince its clients the concept of Enterprise in a Box. The concept of a fully loaded rack with enormous computing capacity and all relevant Enterprise Applications pre-installed and ready to go with minimum fuss is at the core of this offering, and this latest offering is a logical extension of that trend. While the OEEC does make a compelling proposition on the face of it, its not really a Cloud-In-A-Box as Oracle would like us to believe. The way I see it, it's a highly potent platform which supports dynamic resource allocation and metering.
Lets look at the things that work in this offering :
- Faster than the Cloud : Definitely. With a super optimized platform and co-located software (yes, it all resides on a single physical box), the latency is obviously minimal - something which is not realizable over a network in a real multi-tenant solution such a remotely hosted Cloud.
- More Reliable than the Cloud : Lets face it, Security has been a top inhibitor for Cloud adoption, and with OEEC, these concerns are greatly reduced due to, again, the 'co-location' factor of the OEEC offering. Its all still within the Enterprise. Its also more stable, at least for the on-board components due to the robust architecture incorporated.
- Greater InterOperability : The Software stack is all developed on the Fusion platform and once the integration of all the acquisitions is complete, the software stack will be seamless and highly efficient. The enhanced compatibility of the software lends itself to simpler configurations and thus it takes much less time (and cost) to get it up and running. In fact, the software deployed to the reference system uses default tuning and configuration to get you started. The platform is also closely integrated with the Oracle Cloud platform.
Why will it not replace the Cloud ?
- Higher Costs : Well, lets face it, this just another fancy hardware(plus software) platform that needs to be purchased up front and installed before you can start using it. While, the exact cost details are still not known, there is definitely a bulk investment required. The usage metering is to allow charge-back from the users as opposed to metering the TCO for the enterprise. In that sense, its no different from setting up a traditional server farm for hosting your applications.
- No genuine Multi-tenancy : The enterprise applications are all hosted on the same machine and if the machine is decapitated for any reason, the only resort is to fall back to the DR option, which is no different from what needs to be done in a traditional "Server Cluster" world.
- Still Requires Maintenance : One of the most compelling offering of Cloud Computing is the almost complete lack of any maintenance requirements for the users. This is in-fact a basic tenet of a service. It should be noted that the almost 60% of the TCO during the life-cycle of a platform is contributed by maintenance costs and its this cost which is obviated by the Cloud Service. The OEEC still requires a dedicated platform support team to keep it up and running.
All said and done, the fact that the Enterprise should concentrate on running its core business and not spent effort (and mind share) on the platform is one cogent argument offered by the proponents of the Cloud Computing mantra and hence this offering from the Red Giant still doesn't qualify as a real "Cloud in A Box".
For the moment, I am with Benioff, a former Oracle executive and now an Oracle competitor. "Clouds aren't in a box," he said during an earlier OpenWorld session. "That's the whole idea."
Comments
Nice post. It raises a number of very interesting points. Vendor positioning, whilst entertaining can often confuse decision makers. As cloud is still very much at the formative stage of adoption the whole topic is ripe for confusion.
Many companies are hedging their bets and are looking for an 'on ramp' when it comes to cloud – either through private and/or hybrid approaches. A mixed approach to on-premise and off-premise IT is attracting the most attention it seems (at least from my conversations with analysts).
Whilst the dynamic ‘pure play’ vision of cloud, i.e. IT ‘on tap’ as a dynamic resource is compelling, we're not there yet by a long way. The reality is that we're very much on the first steps of the cloud journey. It is not a co-incidence that most (but by no means all) of cloud usage remains confined to the application layer.
Vendors have an important part to play in setting the parameters of the debate, but it is always a great sanity check to refer back to what customers are actually doing today. For most large enterprises cloud will involve a heavy degree of integration – either due to legacy investments or because they have to utilise multiple vendors for various lines of the business. We will see pure play approaches to cloud – but the short to midterm future will focus on private and hybrid approaches.
As to which vendors that will favour depends on who has the best answers for customer’s real world challenges. It’ll certainly be interesting to see how where the market moves to over the next few years. Interesting times...
Posted by: Tris Clark | October 8, 2010 12:34 PM