Tuesday, October 5, 2010

Larry Ellison's view of "Cloud Computing"


  1. I was previously a technology consultant and worked on implementing Siebel (one of Larry's products) at Fortune 500 clients.

    I think Larry's joking reflects a cocky nervousness around the long term disruptive nature of the cloud and Oracle's current business model. The reason that Oracle exists today is that there is an excruciatingly high margin traditional tech executives are willing to pay to ensure the privacy and security of their data within their firewall. Moreover, once these large scale enterprise solutions are implemented, there is so much risk to a business by making changes, the switching cost can be impossibly high (let's face it, no one wants to make a decision that stops product from shipping out of the warehouse). This fear and apprehension in turn leads to massive internal overspend on CPU / Storage / Network Infrastructure and an automatic disregard for the on-demand pay as you go elastic resources offered by cloud providers.

    The current mental model with top executives in the Fortune 500 is that in order to scale a business you need in house large scale heavy-duty enterprise software. This network effect is further emphasized by large enterprise software providers. Don't believe me? Just see SAP's latest ad campaign here: http://www.aviatorww.com/case-study-sap.aspx. The CEO of Oakley's buddy, the CEO at Callaway uses SAP - how could Oakley choose differently?

    As tomorrow's large companies start here in Cambridge as four person MIT start-ups, I don't think this mental model will persist. I think that growing up without the SAPs and Oracles will make it much easier for companies that grow large (a la Akamai) to continue with cloud solutions like Netsuite (for ERP) and Salesforce (for CRM).

    Lastly, on a personal note. Siebel made a "push" into cloud / object oriented packages about 3-4 years ago. In my five years of professional services, I never heard of a single client adopting these tools. They always chose either to remain over-paying for infrastructure and implementing in-house Siebel or going with Salesforce.

  2. I'd like to echo Jarrod's explanation of Larry Ellison's jokes about cloud computing. The jokes center around some over-optimistic statements, for example, cloud computing is revolutionary, rather than building on a pretty objective basis. It reflects his nervousness around the long term disruptive nature of cloud computing, in my eyes.

    I used to work for a pre-IPO company, taking care of group-level sales force management and client resource management. When turning to IT solutions for better efficiency, I was faced with the decision between investing own company’s IT infrastructure to build and run the CRM system by ourselves, or leverage Saleforce’s service. We debated around security, consistency as well, and finally Salesforce won. Investment return mattered! If we bought the equipment, at least 90% capacity were to be wasted, given that we were not going to be a “cloud” by ourselves. We were pretty happy about Salesforce experience. Not sure about large corporate, but cloud computing definitely appeals to small-size businesses.

    However, talking about Amazon, my primary concern is consistency of service, as Amazon’s major focus is still its online retail business, different from the model of Salesforce.com. Users of AWS are inherently put in a interior position, especially during winter peak time. What will Amazon do if in winter its own online retail business requires support of all its computing capacity? A potential way may be that Amazon buys additionally required computing capacity from other vendors during its winter peak time, to ensure service to AWS customers. I am not sure whether Amazon is doing this, but I were a customer of AWS, I would take this into consideration when making the decision.

  3. The previous comments make very strong arguments regarding the threat of cloud computing on large software companies. It is definitely a concern for small to medium sized companies.

    For larger companies however, although I think cloud would eventually attract them as well, I think there will still be a significant number who stay with the in-house model.

    I was working for IBM Global Services for years, and during that time, I don't think any of the clients I worked with would switch to cloud. Privacy is a big concern, not just to the sensitive applications but also to the many connected application that integrate into the data provided by the different systems in the company.

    Another perspective to this arguments would be that the cloud would help companies like Oracle, because it allows small to medium sized companies build application they could not have otherwise. When these applications mature to the point they need a dedicated in-house systems, they would then approach the Oracles and IBMs.

    This being the case, I agree with all previous conclusions that his "joking reflects a cocky nervousness."

  4. So while I think he was a little overboard, I do totally agree with Larry Ellison's notion that the term "Cloud" is over used.

    There is absolutely a new business model for providing IT services that has emerged based on technology change in the last few years. That being said, the term "cloud" is over used and as such has become a cultural buzzword that has fed its misunderstanding. (Hmmm System Dynamics reinforcing loop anyone?)