As details of the second round of Microsoft layoffs have trickled out, the only obvious pattern has been cutting back on what's not core to the platforms and productivity focus that CEO Satya Nadella has been emphasizing (or the explicitly protected Xbox side of the house), plus further tidying up of which teams sit where.
Microsoft Research Silicon Valley
The announcement that the Microsoft Research lab in Silicon Valley was closing is sad more because it ends a piece of Silicon Valley history -- and because talented researchers suddenly have to find new positions -- than because of large areas of key research being lost.
Back in 1984, DEC founded its Systems Research Center in Palo Alto when Robert Taylor left Xerox PARC; in 1996 current MSR Silicon Valley head Roy Levin took over. When Compaq bought December in 1988, the lab changed its name and kept working. But when HP bought Compaq in 2002, although the Systems Research Center merged into HP Labs, Levin took his team of distributed systems researchers and co-founded the MSR lab to give them a home. That included the trio of Turing Award winners Butler Lampson, Chuck Thacker, and Leslie Lamport.
Those are names to conjure with in the history of computing, and their work at MSR has continued to be important for Microsoft. Thacker worked on both the first computer with a mouse (the Xerox Alto) and the Tablet PC team that introduced mainstream pen computing. But some of them have already moved on to other MSR labs (Lampson is based in Cambridge, MA) and they're perhaps more to be thought of as the elder statesmen of research.
The lab was already down to 50 people from a peak of 80 researchers in 2001 and the other programs based on the Silicon Valley campus -- including the StorSimple team, BizSpark incubation, the corporate citizenship program, and outreach to the local technology community by providing free facilities for meetings and events -- remain.
That presence in Silicon Valley had been an advantage for the lab, in terms of tapping university talent from Stanford and Berkeley, and getting researchers who simply wouldn't leave the area. But although the research teams working on concurrency and distributed systems produced interesting tools and techniques, and the list of products incorporating research from MSR Silicon Valley rages from spam-fighting tools for Hotmail to Kinect body tracking to the ASLR security protection that provides key malware protection in Windows, those are all older projects.
Some key projects never turned into commercial success. Dryad, Microsoft's system for graph-based distributed database programming, didn't get the usage Hadoop did and industry focus has moved from server clusters and parallel computing to cloud services. (Dryad and Dryad LINQ are both open source projects now.) The Singularity research OS team came up with some excellent techniques for creating more reliable systems, but that research finished some years ago.
The new Azure DocumentDB service does take ideas like tunable consistency from Pileus, a system Doug Terry's team built to show how you can have fine-grained control and service level agreements even in cloud-scale, geo-replicated storage. Since Terry's biography on the MSR site refers to him in the past tense, it looks as if won't carry on working with Microsoft after the closure of the lab. MSR Silicon Valley was an obvious place to save money, if a rather tone-deaf move by Microsoft when it's still often seen as irrelevant in the area.
Dropping the robotics research team also makes painful sense. They weren't working on robotic systems like self-balancing robots or self-driving cars; they concentrated on research into development tools for robotics, like Microsoft Robotics Developer Studio, and on interactions between humans and robotics.
Both are interesting and important areas but neither of them are going to give Microsoft major market share or breakout success in a market that is still largely confined to academics and hobbyists. The big successes in robotics will go to the companies that build robots rather than the companies that build tools for building robots. This is clearly Microsoft making a painful but practical decision to narrow its focus from working in areas that are unusual, interesting, full of potential, and unlikely ever to have a major impact, to projects where it can really stand out.
What about Trustworthy Computing?
The closure of the Trustworthy Computing Group was initially far more concerning, especially as security isn't an area where Microsoft can afford to lose any talent. When Bill Gates took the unprecedented step of stopping all product development for a month while every developer at Microsoft retrained in writing secure code, it didn't just rescue Windows and Office from the swamp of viruses they were slowly drowning in. It also helped change the mindset across the industry to include the idea that security matters in product development, in design, in deployment, and at every stage of technology. Microsoft didn't just teach its own developers security thinking. It didn't just recruit security experts and researchers to work with the Microsoft development teams, in house or as consultants. It also invited in hackers and penetration testers to tell Microsoft what it was doing wrong, and everything it learned from them, it put into white papers and annual reports and training materials that it made freely available to anyone who wanted to create more secure products. Without the Trustworthy Computing Initiative at Microsoft, we probably wouldn't have an ISO standard in secure development today. If you still think that Windows is an OS with security problems, you haven't been keeping up with the major improvements Microsoft has made, release by release and month by month.
So the thought that Microsoft would back away from that commitment was worrying. Microsoft's John Lambert took to Twitter to reassured people that "all your TwC is still here. SDL, operational security, pentest, MSRC, Bluehat are just under a new roof." The other parts of the Trustworthy Computing Group that looked after privacy and policy found a new home under general counsel Brad Smith, the Microsoft lawyer who has been leading Microsoft's privacy campaign in courts and at legal events around the world.
That division actually makes sense, but like the rest of the Microsoft layoffs it hasn't been well communicated. Employment law means a company has to tell the employees who are being laid off first, and explaining layoffs in detail is one way of inviting commenters to nitpick every single decision, but it wouldn't hurt for Microsoft to explain how these layoffs fit its priorities.
In the long run, these are probably all sensible decisions for Microsoft, which has often taken far too scattergun an approach to what it works on. With research teams around the world working on everything from voice recognition to quantum computing, Microsoft is never going to have a narrow focus. Painful as the layoffs are, especially to the people who no longer have a job, they do come mostly in areas that just aren't that important for Microsoft's future.
This story, "The second phase of Microsoft's layoffs is all about focus" was originally published by CITEworld.