Day 2 at TechReady brought a keynote by Bill Gates and another Q&A session afterwards. The keynote was pretty interesting, it was a fairly typical BillG presentation with some good demos. He highlighted six major areas Microsoft is focusing on over the next 5 years. No real news there, it was the same things we've been talking about publicly for a while, the Services platform, etc. What struck me most about both his keynote and Kevin Turner's is the much increased focus on the consumer space.
Consumers are driving a big portion of the tech industry today and that makes its way eventually into the corporate market. Both are extremely important to Microsoft for obvious reasons. They hype and success of the iPhone, Wii, etc. demonstrate two areas BillG says Microsoft needs to improve. The first is the design and approachability of the hardware itself. The second is the user interface experience like the touch capability of the iPhone and the Wii controller. In each of these areas Microsoft has had research and demos for a long time (like Microsoft Surface) but haven't gotten them to the market quickly enough. My impression was that across the board there will be focus on hardware and user experience.
For the remainder of Day 2 I attended several sessions on Virtualization. The first dealt with using Microsoft technology to create a dynamic data center. The second was about current large scale virtualization projects currently underway. Some things demoed that I haven't had time to test out yet are a Self Service Portal for managing virtual machines. The scenarios are developers needing a virtual environment for testing so they go to the SCVMM portal and provision for themselves a set of VM such as a DB VM, an IIS VM, etc. This scenario can also be used for production VMs where a Tier 2 admin may be givne permissions to create new production VMs. Several TAP customers are doing this in production today on VS 2005 and SCVMM. The second one was the capability of Data Protection Manager to back up running VMs using VSS. If needed, you can snapshot running VMs as frequently as every 15 minutes. SInce it is VSS aware, the VMs and VSS aware applications are properly quiesced for a consistent snapshot.
This is an area I focus on a lot but seeing demos and real world case studies of customers using the full suite really demonstrated the breadth of our offering. Most people think only of Virtual Server when they think Microsoft virtualization but the full suite is quite a bit more including Virtual Server, Virtual PC, Softgrid, Virtual Machine Manager, Data Protection Manager, Operations Manager. The full suite gives us server, workstation, and application virtualization, high availability, quick migration, P2V, backup/DR, advanced monitoring and much more. The suite is a more holistic offering than our competitors and once Windows Server Virtualization comes out, I think Microsoft will gain much more traction in the market as we will have a suite that is equal to or better than our competitors in almost all areas. Certainly a great segment of the industry to focus on.
Check back tomorrow for the Day 3 summary, I'll be attending more Virtualization sessions as well as Softgrid and Configuration Manager sessions.
Day 3 turned out to be even more interesting than I thought. There were keynotes again this morning, but they were very demo heavy which is always good. The first was Kevin Schofield from Microsoft Research. After some great stats about Microsoft being the largest spender on pure R&D and being a leader in annual patents granted, he showed some amazing demos of visualization and image processing projects our researchers are working on. One of them you can see as well at this link. After this, Jim Allchin and some of his team did some great demos of Vista and Office 2007.
Later in the day I went to a great session on Active Directory in Longhorn. The biggest changes are targetted toward branch office scenarios with the introduction of the read only domain controller (RODC). As hopefully most people know, physical security of domain controllers is a critical requirement. Typically in branch office scenarios this is not the case. With the RODC, even if it is compromised, there is not a full account database on the server and it is prohibited from replicating back to full domain controllers. This role can run on Longhorn Server Core so you can have a very secure branch office directory server.
Finally, at the end of the day I went to a session on System Management Server, soon to be called System Center Configuration Manager. The session focused mostly on OS deployment and Network Access Protection (NAP) which were the two biggest areas of investment. Finally we will have a unified and robust method for deploying both client and server operating systems. There are some great improvements to the image based deployment process. In terms of NAP, this is going to be one of the biggest areas to improve security. There are many options for checking, quarantining, and remediating non-compliant systems. It's also extensible so ISVs and even customers can create their own health checks.
Day 3 continued the themes of security and manageability across the board.
It is very cool application virtualization technology. Check out this post for details:
This is the first in what will be a series of posts from the team so be sure to subscribe to their feed. There is a lot of development in this area that I heard about two weeks ago at TechReady like further improvements to the core technology, sorting out some overlapping functionality with other products, and further integrating it with System Center.
By now most readers of this blog will have heard that some of the planned Windows Server Virtualization (Viridian) features have been deferred. The details were posted here. Dan Kusnetzky over on ZDnet has a blog post with good analysis of the options. This was certainly a disappointment for those of us who are excited about the technology. Clearly some work needs to be done on the planning and expectation setting process. The bottom line in a situation like this is that this technology is complicated and sometimes issues crop up that are not possible to anticipate and sometimes take longer than expected to solve. This leads to the situation where you are behind schedule. At this point, as both the orginal post and Kusnetzky's post point out, in software dev you have three choices: Reduce quality and testing to ship on time; Push out the release date for the full product; Cut features and ship on schedule. You then need to make the decision as to which course to take and let your customers know as soon as possible. Given the state of the market, my view is that the team made the right decision.
The decision-making process for what to virtualize should revolve around the resources required by the workload to be virtualized (processor, ram, I/O, etc). Currently, Microsoft virtual machines are limited to a single processor core, just under 4 GB of RAM, and 4 NICs. Consequently, no single server workload that requires more than this can be virtualized without changes such as scaling it out to multiple VMs etc. Windows Server Virtualization is being designed to remove those limitations as a top priority. There are additional priorities such as high availability, dynamic data center scenarios, etc. The identification and prioritization of these scenarios is critical and customer driven as it should be.
Those priorities were planned to be accomplished in WSV via the following major features.
As you can see from the announcement, Live Migration and Hot Add are being deferred. That still leaves several major features (in addition to a lot of other features) still included in the initial release. The features that are in address the biggest limitation with our current offerings, namely the size of the single server workload that can be virtualized. The team in my view is correctly evaluating the market, our offerings, competitor's offerings, and customer needs. I certainly wish all five features would be in the first release but I also would certainly rather have the first three on schedule than to have to wait 3, 6, or more months to get any of them. Put simply, the cup is 3/5ths full...
The first part of a four-part article series titled Architecting a Microsoft Private Cloud, by myself and my colleague Adam Fazio (@adamfazio) has been published in TechNet Magazine here:
Microsoft Services has designed, built and implemented a Private Cloud/IaaS reference implementation using Windows Server, Hyper-V and System Center. Our goal throughout this four-part series will be to show how you can integrate and deploy each of the component products as a solution while providing the essential cloud attributes such as elasticity, resource pooling and self-service.
In this first article, we’ll define Private Cloud/IaaS, describe the cloud attributes and datacenter design principles used as requirements, then detail the reference architecture created to meet those requirements. In parts two and three, we’ll describe the detailed design of the reference architecture, each of the layers and products contained within, as well as the process and workflow automation. Finally, in part four we’ll describe the deployment automation created using the Microsoft Deployment Toolkit and Hydration Framework for consistent and repeatable implementations.
For more information about the architecture, see the Hyper-V Cloud Deployment Guides here (also co-authored by Adam and I) and the MCS Datacenter Services Solution here. Stay tuned for the remaining three articles...