Full disclosure: I work for IBM, a company that competes with Oracle in many areas. This blog states my opinions, my alone, and does not reflect the opinions of my employer, nor do I know what these opinions may be. I hope that while you keep my bias in mind you will evaluate my arguments on their merit.
About this time last year Larry Ellison did his famous anti-cloud rant. A year has passed and two weeks ago the headlines were “Oracle OpenWorld is all about Cloud Computing”. Has the outspoken and flamboyant CEO changed his mind about the cloud? Is this a signal of a complete reversal in Oracle’s Cloud Computing strategy? Peering a bit behind the showmanship, it is quite clear that not much has changed and despite tagging everything company releases as “Cloud Computing” Oracle is still very troubled and afraid of the advance of cloud computing; justifiably so. Larry Ellison’s rants against the cloud have not been all idle banter. Shortly after it acquired SUN, Oracle killed SUN’s newly announced cloud; Oracle was not going to be in the cloud business. Why not?
At the top of the cloud, Oracle is at the bottom of the food chain
There is no arguing that Larry Ellison has been brilliant in leveraging Oracle’s market share in DBMS to build an integrated hardware+middleware+applications powerhouse. He has taken the page right out of the IBM playbook and established Oracle as a top tier vendor for world’s largest enterprises. The old IT expression “you don’t get fired for buying IBM” can now be applied to Oracle. But now, with the advent of cloud computing, Oracle is facing a serious issue. These large enterprises have developed quite the taste for consuming applications as services delivered over the internet. This is in preference to having IT departments struggle for years to plan, budget, provision and run application software in their own data centers. The industry term for this phenomenon is SaaS (Software as a Service). And when it comes to SaaS, the balance of power has shifted from IT to the line of business and from IT vendors to the fledging SaaS providers. IT is still in the loop but it is no longer the sole decision maker, often not even a key one. This presents not an insignificant problem for Oracle, who’s sales people do not enjoy the same relationship with the line of business executives as they have cultivated with IT brass. To make matters worse, when it comes to SaaS, Oracle is not a category leader not even the top tier vendor; it is not even on the radar screen. Mention SaaS and Salesforce.com is going to be the first name that comes to mind, not Oracle/Siebel CRM suite for example. It is not that Siebel (and after its acquisition Oracle) has not tried; it just failed miserably. Oracle CRM on Demand is generally considered a lower tier SaaS CRM solution and is not on the top of consideration list. For a company that is built its success being a category leader this is a very uncomfortable place to be and a very scary proposition. Is Oracle the only enterprise application vendor caught in this predicament? No, not by a long shot. However, being on a sinking ship with a bunch of people you’ve been locked in a mortal combat for years is not a great place to be. So, what is Oracle to do? Short of going on another acquisition spree, Oracle has no choice but to try to keep this sinking ship afloat by discrediting the validity of SaaS and co-opting the definition of Cloud Computing to suit its needs.
Low margin high volume business is not for Oracle
Is it any wonder that Saleforce.com became enemy number one at this Oracle OpenWorld? Larry Ellison quickly set aside HP and Mark Hurd melodrama and zeroed in on his former protégée and a new arch rival Marc Benioff and SalesForce.com. Larry’s most persuasive argument against Salesforce.com is “all the data is mixed up together”. Larry Ellison is a very smart man completely on top of today’s technology. Surely he knows the right IT term for “all mixed together” is “multi-tenancy” and it is precisely this sharing of resources through the multi-tenancy that is the secret sauce that SaaS and IaaS (Infrastructure as a Service) vendors deliver real cost savings. It allows software provider deliver services to the entire user base out of a single instance of the application (or at least a small number of instances) enabling never before achieved levels of resource sharing. For example, Saleforce.com serves hundreds of thousands of users a day using 16 instances of its application. As a result, SaaS customers are able to enjoy services at a fraction of it would cost if these services were delivered in a dedicated fashion. By running a single copy of the application SaaS vendors are able to not only greatly reduce the costs for their customers but also to accelerate the pace of innovation. After all, they don’t have to spend resources maintaining ancient versions of their software … every customer is running the latest version; the only version. Cloud providers like Amazon, IBM, RackSpace, Teramark and others rely on multi-tenancy at the hardware level to share compute, storage and network capacity across all of their clients. It is precisely this sharing that let’s them rent compute capacity for as little as 2 cents per hour. Just as a point of comparison, it costs me about 5 cents in electricity alone to run a server in my house. The reason Amazon is able to offer such low prices is because it can share the costs across its entire customer base and I don’t share my server and costs with anyone (my kids refuse to pay their fair share). This is bad news for Oracle. Larry Ellison claims that Oracle’s vision of the cloud is more like Amazon EC2 than Salesforce.com. Reality though that what Oracle is delivering is nothing like Amazon cloud. Oracle just announced Exalogic system as “cloud-in-a-box” with prices starting at a cool $1 million and growing in $1 million increments. You can rent a whole lot of hours from Amazon at $0.02/hour before that $1 million investment becomes attractive. Sure, if this is an enterprise data center purchase, you could share its resources between multiple departments and even divisions but you still need to make a very large capital investment up front. This is something that Oracle very much likes as the “pay as you go” model of the cloud is terribly scary if you are addicted your customers making large purchases. I say “addicted” because Oracle more than any other enterprise IT vendor has built its channel around very large, very well paid, very aggressive direct sales force. The compensation structure and the entire channel is built around selling very high margin products to a relatively small set of very large enterprise customers. When you are dealing high margin large deal size enterprise agreements, you can afford gold plated sales force, golf tournaments and trips to watch Larry win America’s Cup (congratulation on that). On the pother hand, if you want to get an account on Amazon, Google or Salesforce.com, you don’t get to talk to a sales person; just a web page where you can enter your credit card number. This is very low margin high volume business; business Oracle can’t compete in. What will Oracle do? It will mount a campaign of FUD trying to scare customers away from using any kind of shared services. Oracle will tell customers about Salesforce “mixing the data all together” and will produce videos of fictitious customers (just like they did at Oracle OpenWorld) that are bringing apocalyptical consequences on their business by going to the cloud. Oracle will tell us, just like he told Marc Banioff (watch the video) that “on-premises” systems is the right, no, the only solution. It is indeed for Oracle as that is what it sells and that is where its high margin business is.
“Pay only for what you use” denies Oracle opportunity to oversell
Wait a minute, you may say. Didn’t Oracle just announce that you can run Oracle applications on the Amazon EC2 cloud. Yeap, you can, but only if you purchase these applications from your Oracle sales person up front. Is that bad? No, not at all. For many customers this is the right model. But where the cloud is different from traditional enterprise consumption is in its “elasticity” and associated “pay only for what you use” model. The elasticity of the cloud is what allows you, the customer, to avoid large upfront capital expenditure or as cloud people say “no CAPEX”. It allows you to have the capacity you may need on a moments notice and without over provisioning or reserving it upfront. The best part, it allows you to stop using and stop paying for the capacity when you no longer need it. For a vendor who build the business by over-selling capacity to cover peak workloads and then some, this is indeed a very scary future. This spells much smaller deal sizes (if there is even a deal at all) and it means much higher customer churn. After all, if you are paying monthly or by the hour, there is no guarantee at all that you will come back next month. You have very little vested and are free to pick a new provider any time you choose. Oh, the horror … no multi-year multi-million $ Enterprise License Agreements. No more mandatory 20%+ Service and Support fees. No more license compliance audits to nail your sales quota. If you are an Oracle sales rep, this is horror indeed.
So, Oracle is afraid of little clouds …
Oracle has a lot to loose if, or more precisely when, Cloud Computing goes mainstream. Oracle can’t rely on product leadership to impress your line of business executives. Your Oracle salesperson will be singing you the sweet hymns of Exalogic and “cloud in the box”. She will be telling you how you should get it in to the ELA that she is working on. I am not saying that that big box with an X on it is not an impressive piece of hardware. However, if your project calls for “Cloud Computing” you should look elsewhere. Beware of new prophets with a hastily developed vision of the cloud, the one that is no cloud at all. It may have “Elastic Cloud” written on the outside but it stretches your budget $1 million at a time and it never shrinks. It will make you pay upfront and will have your Oracle sales person come back every year to get their customary 22% renewal fee. And you that old saying “nobody ever got fired for buying Oracle”, well, when it comes to the cloud the jury is still out on that.
I said at the beginning that this post is my opinions. You may have a similar or a different opinion. I’d like to get comments from those who share my views, but I would love it even more if those of you who disagree would post your comments. I do moderate comments on this blog to filter out spam and really offensive language but I never sensor. So, what are you waiting for, scroll down to the comments box and let the world know your thoughts.