HP's Joe Weinman delivers keynote at Tier1's Hosting & Cloud Transformation Summit 2011
(WEB HOST INDUSTRY REVIEW) — Joe Weinman (http://www.joeweinman.com) of HP delivered the third Monday Morning keynotes at HCTS, offering launching into an interesting talk that turned a lot of common notions about the cloud business on its head.
He delivered them kind of point form, so I’ll try to cover a lot of the ideas that way:
- Cloud is not a brand new idea, or a new business model.
- It is not just services accessed over the web, via a browser
- It is not just pay-per-use or on-demand
- Large clouds don’t necessarily have huge economies of scale or price advantages
- Economies of scale are not the key advantage to key services.
- The market is not going to be limited to just a few big players
- It isn’t necessarily important to replace CapEx with OpEx.
- Not all rational decision makers will want to save money
- The cloud will not replace all enterprise data centers.
One of his key hopes for the session, he says, is that the audience walks away with some questions about the way we think about the cloud.
He tries to quantify “value” in the cloud, with key factors being cost reduction for units, total cost, opportunities, improvements in time and profitability, growth in marketshare and revenue growth, customer experience, customer satisfaction, and many others.
It’s difficult to capture his deadpan in text, but it’s probably pretty important to point out that Weinman somehow managed to combine a very funny deadpan with the understanding that he’s very serious about the economies around the cloud.
On the issue of seriousness, he points out that there are a lot of equations on his website that serve to support his main assertions here, and he would steer anyone who disagrees toward that site in order to better understand the foundation for what he is saying.
It is very significant, he says, that cloud technologies are making the case that the unit cost savings between a service provider and a cloud service provider. There is data to support the idea that there is a cost savings in running enterprise IT versus outsourcing.
A more compelling argument for the outsourced cloud service, says Weinman, is demand variability. Only a small percent of web properties have more or less flat day-to-day traffic to their applications. The majority of sites have significant variability and unpredictability.
In designing a solution to accommodate demand variation, it is important to consider the peak demand, the average demand, the duration of the peak and the area of the peak demand.
Following a fairly inscrutable slide of equations, he says that, all other things being equal, if cloud services cost less than enterprise IT, then use them. If they cost more, then don’t jump to conclusions, because if demand is “spikier” than the cloud is costly, a pure cloud solution will cost less than a dedicated one.
He also says, (again, all other things being equal) a hybrid solution is almost always best. And that will evolve into a future model where “hosting” and “cloud” will be two different consumption models for capacity that differ based on commitment, which allows the service provider to guarantee more utilization, and therefore provide long-term services at a cheaper rate.
He goes very quickly through the idea of aggregating the cost of traffic spikes, but the point is basically that aggregating the traffic spikes of many customers sort of normalizes the network costs associated with those spikes, for service providers. More significantly, though, the cost savings come fairly quickly and flatten out fairly quickly, meaning that a medium-sized provider’s savings are very close to those of even the biggest provider possible.
No related posts.











