I’ve been getting a number of inquiries recently regarding benchmarking potential savings from consolidating multiple physical servers onto a smaller number of servers using VMs, usually VMware. The variations in the complexity of the existing versus new infrastructures, operating environments, and applications under consideration make it impossible to come up with consistent rules of thumb, and in most cases, also make it very difficult to predict with any accuracy what the final outcome will be absent a very tedious modeling exercise.
However, the major variables that influence the puzzle remain relatively constant, giving us the ability to at least set out a framework to help analyze potential consolidation projects. This list usually includes:
As someone who has been covering cloud computing since the dawn of Amazon’s Elastic Compute Cloud (EC2) I’m constantly in education mode about what is and isn’t cloud computing. To borrow an analogy from my Forrester colleague Ted Schadler’s keynote at last year’s IT Forum, the challenge is a lot like helping blind men discern an elephant through just the parts of the animal they can reach. One feels the trunk and declares it a cylindrical, yet hairy and warm snake. The other calls it a strong, tough and deeply rooted tree upon feeling its hind leg. Each examiner brings their own experience and context to the challenge as well as their own judgments, then leaps to the conclusion that best fits their desires.
Today, Google announced Google App Engine for Business, and integration with VMware’s SpringSource offerings. On Monday, we got a preview of the news from David Glazer, Engineering Director at Google, and Jerry Chen, Senior Director Cloud Services at VMware.
For tech industry strategists, this is another step in the development of cloud platform-as-a-service (PaaS). Java Spring developers now have a full platform-as-a-service host offering in Google App Engine for Business, the previously announced VMforce offering from salesforce.com, plus the options of running their own platform and OS stacks on premise or in virtual machines at service providers supporting vCloud Express, such as Terremark.
What’s next? IBM and Oracle have yet to put up full Java PaaS offerings, so I expect that to show up sometime soon – feels late already for them to put up some kind of early developer version. And SAP is also likely to create their own PaaS offering. But it’s not clear if any of them will put the same emphasis on portability and flexible, rich Web-facing apps that Google and VMware are.
So Google aims to expand into enterprise support – but will need more than the planned SQL support, SSL, and SLAs they are adding this year. They'll also need to figure out how to fully integrate into corporate networks, the way that CloudSwitch aims to do.
IBM has been talking a good cloud game for the last year or so. They have clearly demonstrated that they understand what cloud computing is, what customers want from it and have put forth a variety of offerings and engagements to help customers head down this path – mostly through internal cloud and strategic rightsourcing options. But its public cloud efforts, outside of application hosting have been a bit of wait and see. Well the company is clearly getting its act together in the public cloud space with today’s announcement of the Smart Business Development and Test Cloud, a credible public Infrastructure as a Service (IaaS) offering. This new service is an extension of its developerWorks platform and gives its users a virtual environment through which they can assemble, integrate and validate new applications. Pricing on the service is as you would expect from an IaaS offering (and free for a limited time). If you are testing with IBM software you can either bring your licenses or check out the equivalent instances from their service catalog. There’s even a new version of Rational Software Delivery Services for shops familiar with Jazz.
Zimbra has been the sleeper cloud-based email provider for the enterprise. I've known about the Bechtel deal -- roughly 50,000 seats globally -- for some time, but couldn't talk about it. Though it's been a while since I've spoken to Ramesh May, he did share some important facts with me:
1. Zimbra's code base is open source, with a 20,000 active members in the community. The Zimbra code base runs on Linux. It can be downloaded to run on-premises and it also is the foundation of Zimbra's cloud email service.
2. Yahoo! Zimbra was selling an email seat for $28/mailbox/year for 50+ seats. We'll be interested to see how the pricing changes.
3. The company was working with the community on adding instant messaging, expanding widgets, and building an offline email client. We also saw some interesting mashup and document viewing features.
4. Back in April, the company had 130 employees, 600+ .edu customers, 44M mailboxes, and 60,000 customers.
So why hasn't Zimbra been bigger on the national stage selling its hosted (80% of seats) and on-premises (20% of seats) email and calendaring solution? Two reasons.
First, Yahoo! did not build a direct sales force that way Google and every other enterprise email provider did.
Second, because a lot of these seats are sold through service providers. Comcast and NTT Communications have been selling Zimbra seats. You may be running Zimbra and not even know it.
So now it becomes clearer why VMWare bought this massively successful email provider.