WALKING THROUGH SemiCon West, the first booth I ran into was a firm which I simply would not have expected at a semicondutor equipment show, Microsoft.
Yes, that Microsoft. It seems it is fairly large in the semiconductor manufacturing space with a raft of different prodcts. Who would have thought it?
Microsoft is big is this area in two ways, platforms to enable partners to talk, and extending data to the tools you use every day. The platforms are the part I wasn’t really expecting, they are pushing BizTalk Server and Sharepoint as the tools to make a platform that everyone can talk on.
It is one thing to have an API, you can theoretically exchange data with a solid API, but that says nothing about what you run to pass data to that API. Even with that worked out, you still can pass garbage back and forth, but at least you can pass it.
What MS is doing is not middleware, but a platform to build your manufacturing software on. If you go MS from end to end, it solves one of the problems, can you work with the data you get. If you are both MS-centric, and use the same APIs, it makes talking a lot easier. You can still pass garbage back and forth, but you can run that garbage much more easily.
The APIs that MS is using is Rosetta.net, so that part is pretty much as industry standard as you can get. Two out of the three problems solved, the last bit is sending the right data. MS hopes to make that the only thing you need to worry about, your only worry is the things that matter, not the plumbing.
So, does it work? It seems MS is keen on the whole concept of eating their own dogfood, and in this case, they did. MS is now a fabless semiconductor house thanks to the XBox360. The original XBox was more or less off the shelf PC parts, the 360 is almost all custom. MS needed to build a GPU with ATI, a CPU with IBM, and countless other parts with hundreds of other suppliers. Add in second sourcing, and you have the potential for a lot of partners that all need to be in the loop.
The problem at this level of the game is split into two categories, public and private data. Public is the inter-company data exchange, usually governed by the Rosetta.net processes. Security is paramount here, I am sure Sony would just love to know what the yields are on the 360 CPU. All jokes aside, having MS on both ends can simplify the implementation of security protocols, you just use the Windows ones, and they talk to the other end, hopefully always.
The private data is a little more interesting. How do you get the rest of a company to easily grab the information that is hoarded by engineering, research, or those weenies in marketing? If engineering has an app, it probably won’t talk to the one is marketing, and if it does, I am guessing the marketing guys don’t understand CAD and CAM to the level necessary. The flip side of this is if you hand an engineer a 3 page memo on synergistic flows among valued partner ecosystems, they will, with luck hit you with something heavy. If the world in general is lucky, it will mean one less marketeer in the gene pool.
This is where the other thing MS is pushing for, extending the data to tools you use, comes in. MS is now touting Office as the Office System, complete with servers and collaboration tools. The point being that if you are able to pass the engineering data to the server as data instead of as a formatted file, the marketing weenies can open it in Excel. Conversely, they engineers can look at the data without the flowery prose.
The eventual goal is for MS to enable you to look at the entire chain of a product like the XBox360, from the semiconductor design to an individual purchase at Best Buy. The more of the chain you can get in a visible form, the better you can forecast demand . The easier you can make it for a CxO to get that data in a way they can understand, the better they can utilize that information. This is the target that MS is looking to hit, tools familiarity with deep data to back it up.
Did they do it? Not entirely. The 360 was not done completely on MS software, not even close. They obviously tried to get partners to use their software as much as possible, and many did. Others did not, but they still were able to work within the system, and the end result speaks for itself. The 360 beat it’s competition to market by a year, at least, and they were able to ramp volume much quicker than they would have by mere guessing.
Was it perfect? Far from it, if you tried to get a 360 in the first 4 months or so of it’s life, you know first hand some of the problems with supply and demand. Without the manufacturing software infrastructure MS put into place, it could have been much much worse.
News source: THEINQUIRER