Microsoft Making Internet Appliance Chips

M$ Mole writes: "According to CNN, Microsoft is now developing their own chips for WebTV and other new internet appliances. The article is lacking in terms of technical details of the chips, but does bring up a good question of: What does this do to the Wintel relationship?" The idea of Microsoft making chips will raise a lot of eyebrows ceiling high, but it sounds like a fairly modest endeavor thus far, not MS jumping into the ring with AMD, Motorola, Intel, or even with the smaller X86 makers. As M$ Mole and the article say, it's about chips for appliances -- for now.

I've Been Telling People...

...not to invest in software companies. Why? Two words: Free Software.

Free Software is great for hardware companies. It sucks for most software companies. RedHat will never pull in the dough like MS did.

Now, MS is one of the few software companies with the $$$ and wherewithall to transform intself into a hardware company via initiatives such as this, the X-box, and their various PDA efforts.

A lot of other software companies are just going to go *poof*.

The way of the giant

So M$ is building a custom chip to keep the hardware costs down on their low-cost internet appliance. There is a slightly better version of the story on the Mercury News [sjmercury.com].

Lots of companies do this when the cost of assembling a bunch of separate components gets to be too expensive. If you know you have a large market, it is cheaper in the long run to invest in designing a custom chip to perform a single function. It eliminates all the overhead cruft of general purpose computers like the intel architecture. In simple economics terms, this is the easy answer.

For those with a suspicious bent towards anything M$ does, it could be a slap at intel or a first step towards creating a computing platform where competitors can't run. They could be trying to make a system with integrated audio/video streams which will only play a proprietary format which M$ controls, and since the codec is in hardware, no competitor could weasel its way onto the box and steal some content marketshare. Your call.

It'll be interesting if these new boxes turn out like closed architectures, like gaming consoles. Why does that sound like a challenge to figure a way to install Linux? :-)

the AC

MS decision making flowchart

Q:Can we out FUD Palm?
A:Nope. Tried that.

Q:Can we out market Palm?
A:Nope. Tried that.

Q:Can we lock in users on the apps level?
A:Nope. Tried that.

Q:Can we lock in users on the OS level?
A:Nope Tried that.

Q:Can we lock in users on the hardware level?
A:I guess so. We have nothing to loose.

Q:How about giving the customer a better product?

A:Blank stare . . . [laughter]

More at SJ Mercury site

Here is the complete article [sjmercury.com] at San Jose Mercury's site.

You got it all wrong.

Microsoft is now developing their own chips for WebTV

That would be... potato chips.

Check out the register.

The register [theregister.co.uk] has an interesting article [theregister.co.uk] about this.

Particularly, a couple of quotes from Intel about this:

• "I think Intel's reputation as a chip company is better than Microsoft's, and you can take it from there."

  - Ron Smith, a senior VP at Intel's wireless division in Santa Clara

• "I have no problem competing with Microsoft."

  - Mark Christiansen, Intel's senior VP in charge of its IXA project

This may also answer Hemos' question [slashdot.org], about why is Intel demoing Linux failover.

You sell chips: we push other operating systems.

Is this Talisman all over again?

About four years ago the Soft tried to make a revolutionary graphics leap forward with the Talisman chip [microsoft.com]. It was actually a pretty cool design. And like this new chip, Talisman could "take it to the next level", something MS felt it needed to do to make Windows a competitive game platform. It failed for a number of reasons. One of them was the complexity level was higher than any of their fab partners were used to dealing with. Another was that other graphics chip manufacturers became scared to talk to them -- they didn't really want to support Talisman, but felt they needed to get Direct3D support for their chips, and Direct3D and Talisman capabilities were getting intertwined inside MS. The result was a giant mess, and it was finally dropped. My point is that MS doesn't have a very good track record with this sort of thing. Not predicting doom, but I see some similarities between the two.

Networked Appliances

No, you don't need your current appliances networked, but you will want your new appliances to be networked.

I have ReplayTV. I want it networked so that I can log into it from work and see what's recording, delete stuff I don't want, record shows that I forgot to ask it to record, and such. When I watch TV, I want to be able to call up the IMDB page for the movie I just watched.

I want to have my MP3 player networked.

I want my alarm clock/radio to also play MP3s, so I want it networked.

I would like a lot of my house controls (lights, heat, AC, and such) computerized and networked. So I went on vacation and forgot to turn off the AC? I can log in and stop wasting electricity, and program it to be cool again just before I get home.

I would love to have my car networked. It could search for low gas prices on my intended route when the tank gets low. It could report its location if it gets stolen. Obviously, it could download MP3s for the stereo.

I would like to have my doorbell networked. I have a friend that has a doorbell with an intercom, along with a web cam all computerized. Someone can ring the doorbell when he is at work. He can answer on the intercom and look at the person at the door, making them think he's home but can't come to the door.

Ten years ago most people didn't think they needed their computers networked. All it takes is a little imagination. Sure, the value-add may not be that huge at first, but others will imagine a little more, and soon we'll wonder how we ever got by without having everything online.

WebTV has been making chips since the beginning

WebTV has been designing custom chips from the beginning. The founders are old hardware jocks (Perlman and Leak) who did great things at Apple with their cool video TV systems. They were the first that made it possible to watch TV on your mac in a window AND drag that window around. It was way cool at the time. They took this expertise and developed the custom chips for WebTV.

Since then, they've done many revs. Sure Microsoft bought them several years ago, but designing new chips is not new.

This is depressing

How am I gonna crash my Windows box if I can't get it to boot up in the first place?

NSA_KEY

Maybe this will make it easier for MS to put government backdoors in their products. Backdoors in hardware are much easier to conceal and harder to circumvent. Clipper Chip anyone?

Making your own chip

If you're selling over 100,000 units/year, it often makes sense to design your own ASIC.

It's really not a big deal to make your own chip. When I was doing grad school part time several years ago, I made this little chip, together with a small group of other students [pjrc.com]. The whole thing only took a couple months to design. I learned a lot and since then I've had a much better perspective about how ICs are designed, which has been helpful designing at the board level.

The CNN article is remarkable vauge about what Microsoft's chip actually does.... it may be a CPU, or maybe just "glue logic". Whatever it is, it's common to design ASICs for high volume products. Unfortunately, it also common to make a big deal out of nothing.

Convergence

Ignore all of the yoyo's bleating out "but why does my toaster need to be networked" and 1001 variations of trivial-appliance-with-blue-screen-of-death.

This is about MS moving out of the computer and into your TV. Not the good ole rabbit-ears TV, not even your cable-hooked-plus-VCR TV but tomorrows TV.

Think Smart-Cable-box + WebTV + Tivo + Digital Download of Media (music, movies, special events) + Games + Network Sharing + Remote Applications + Home Automation + Telephony.

One box that plugs in, from one vendor, with massive name recognition and tons of back-end architecture already in place. All of your couch-potato needs from one source.

• Smart-Cable-box: Plug it in and it talk to the cable-company. Figures out what the local specs are & automagically configures itself.
• WebTV on Steroids: Browse from your couch - or in a window on the screen, or pop directly to the show's website or just point to the starlets outfit and order it online.
• Tivo: Wanna watch a show later? All of the features of a Tivo/Replay/etc. but from a big name vendor and more heavily integrated.
• Digital Download of Media: Media Player on steroids. Why bother digitizing a program when you can get it already that way? Want to see "Harold & Maude"? Put a request in for it and it'll get downloaded overnight. Pay extra and watch it live. Excited about that special club remix of Brittny Spears? It's on your box for a buck or two. MS has been working on digital delivery for years - this is the terminal.
• Games: Think X-Box light. Think Quake I availiable for a rental fee.
• Network Sharing: Want to plug in your Windows PC? Hook it up and it'll be automagically configured through MS's gateway.
• Remote Applications: Rent MS Money or Word for the evening.
• Home Automation: Want to control your hall light? Buy the MS compliant outlet controller and it's taken care of. MS has been involved in a series of these projects over the years but putting the controller in a smart box that's easily upgraded could be the breakthrough.
• Telphony: Your cable-co already offers tons of 'free' features if you sign up with them for your phone services but they're all the same ones the copper-wire folks offer. How about a universal inbox including your voice-mail? No "push pound-one to..." just point & click. Gramma calls? The TV flashes her name.

  So why a custom chip? Control. Now MS can put all of the anti-piracy / media-control / encryption right into the hardware. Optimize the CPU to run MS architecture material. Heck, with WinHEC they've been setting the specs for years now, it's a small jump to just doing it directly.

  Microsoft doesn't want to be your OS vendor, or your applications vendor, not even your ISP or cable-company or channel - it wants to be all of them.

  Yesterday the MS WebTV, today the MS Phone, tomorrow the MS Information/Entertainment/Shopping system.

  Convergence.

C#, .NET, and more

I'd look at this as a means for Microsoft to bypass the hardware market all together. If they can manufacture and market a WebTV box that uses the .NET infrastructure and the C# language as a development environment, they can bypass Intel, Dell, etc. altogether. And, keep those profit margins up.

You may be able to file this in the "set-top box" file, and safely forget it. This is either a really brilliant move, or a feint to keep the wolves at bay.

Microsofts own processors

Made from 99.9% recycled Intel

Well, maybe not, but in accordence with standard embrace, extend, extinguish philosiphy, I would have to say yes.

But my question is (aside from perhaps the stereo and tv) why does anything in my house besides my computer need to be networked? I don't need web access on my toaster, blender, microwave, refridgerator, washer, or dryer. If you can wire up my sink to automatically rinse the dishes and put them in the dishwasher for me, while having my Mindstorm's clear off the table I just ate from, then *maybe* and only maybe, will I feel that its necessary to have my appliances networked.

soon i'll be surfing the web from my toilet paper spindle

Don't Panic

From reading the article, It's not like Microsoft itself is creating the chips, one of their companies is, webTV. This really isn't that big of a leap for a company to make.

This, of course dosn't mean that M$ isn't the evil empire...
___________________

Re:I've Been Telling People...

Software, however, has massive profit margins, largely due to its different economies of scale. In hardware, like most other manufatured goods, your first few units a very expensive. Then, as quantities go up, your costs go down -- that's the benefit of mass production. However, once you reach a certain critical point, finding the resources and capital to make more actually starts cutting back into your profits -- i.e., when Intel is already running a peak capacity in their PIII fabs, having demand shoot up 25% in a month really wouldn't be a good thing, becuase they'd either have to miss ship dates or throw a bunch of money at farming out or ramping up their capacity.

Software, on the other hand, almost never hits that saturation point. 95% or more of the cost of making a program is incurred before the first copy even ships: development, marketing, testing, etc. Once copies are being boxeed and shipped in large numbers, each one only costs the company an additional few cents for duplication, printing, and distribution.

Now, enter the Free Software movement (or at least its popular media recognition): you can get your OS, server applications, and business tools absolutely free, with the source code, on your choice of hardware. Connected by the Internet, thousands upon thousands of developers toil away on labors of love, making their OSS projects into the best tools on the market.

One might think that this spelled disaster for the old-school software houses, who relied on a steady stream of income from every shrink-wrapped box. However, that same Internet that made the spread of quality, free tools possible also makes possible a new kind of company...the ASP. ASPs have many of the advantages of the software industry: low cost per unit, easy distribution, etc. However, it also allows for new levels of user authentication (preventing piracy), planned obsolescence (you can only buy a subscription to a service, and the ASP changes the software at will), and lock-in (once all your corporate data is on another company's servers, you're going to think twice about telling them to go screw themselves).

Microsoft, as the world's largest developer of new software, is uniquely positioned to take over the ASP market. They can do this either by moving Windows, Office, and the rest of their end user applications to an ASP model, or by working to become the "standard" developers of ASP platform development tools and applications. With personal hardware thrown into their stable, they can insure that every WebTV box, PocketPC PDA, and X-Box console speaks the Microsoft dialect of networking, and reads and writes exclusively Microsoft documents.

.NET is Microsoft's ASP power-play. If C#, DCOM, et. al. can become standards for server-side distributed business logic, then anything that doesn't play nice with them runs the risk of becoming very unpopular. This is why the success of Linux and *BSD on the desktop is a noble, but less important goal -- the battle now is for control of the network, and the network will be the "killer application" for many years to come.

Jeez.

The Solo2 chip -- named after Solo, WebTV founder Steve Perlman's dog

I think that should be nominated for stupidest name yet. It would have been alright if they had called it the Han Solo2 or something. Jeez, even chewbacca is a better chip name then that.