MOST seems to be a de facto standard for what it is trying to do. It wouldn't have gotten there if car manufacturers weren't on board. If I'm a car manufacturer, I want the cheapest solution that *works* for the problem I want to solve.
Ethernet is proven, low-cost, and well understood, but it won't work for *every* networking application, and this seems to be an area where ethernet won't do. If it did, why isn't the auto industry using it?
For Microchip, the issue is the same as it is for any company venturing into a new area. How good is the opportunity? Do they *understand* the market? Do they recognize how the business and sales cycles differ from the ones they are accustomed to dealing with, and what changes they will have to make in how they do things to be able to address that market?
One of the original questions raised was the difference in business models of Microchip and SMSC, and how compatible they were. A company turning out scads of low cost chips for a big market and relying on volume is a different animal from a company making a lower volume of higher priced chips for more specialized markets. They fact that they are different underlies some of the questions regarding the merger. How well does Microchip understand the markets SMSC addresses? How well does it understand the differences between its business model and SMSCs? How well will it do at blending them, and executing equally well in those different markets?
Some classic failures have come about as a result of acquisitions where the acquirer *didn't* understand the differences, tried to manage the acquisition the same way it managed itself, and went belly up in consequence.
We'll see. I see opportunities in the auto market for Microchip if they have that understanding, and can use SMSC's established presence to address those opportunities.
Well argued points @DMcCunney...if I am a chip maker I would gladly support a new communication protocol like MOST as it is a new market for me, and yes it is a big market...but if I am a car maker I would want to have the cheapest, proven technology so I should vote for Ethernet not MOST...but perhaps Ethernet doesn't cut it technically, maybe there is a media aspect of this communication car technology that is served much better by MOST...Kris
Ethernet may be larger in terms of devices shipped, but it's not the only network architecture.
If I'm a chip maker, I'm concerned with volumes. If I make a chip supporting a certain kind of network, is there a big enough market for it? How many units can I sell? How much must my price be at that volume to make money? Can I get that proce?
Every car is now a smartcar, with it's own local area network (likely based on CAN, first developed at Bosch) and chips passing messages back and forth across that network. Devices that interact with the driver and passengers are joining the mix, and MOST appears to be an attempt to aid that integration.
In terms of Microchip and SMSC, the automotive market is a very large one, and design wins there will be attractive targets. Since the cost of chips aimed at that market will be a far tinier percentage of the total cost of building a vehicle than the corresponding cost of, say, a chip in a smartphone, I would expect to be able to charge more per chip. It may not be a low margin commodity business.
thank you @DMcCunney...interesting link...I guess MOST protocol addresses the media aspect more than other technologies...but I still think this is not the most efficient option in terms of economy, Ethernet is bound to be 100x or 1000x larger in terms of devices being shipped than MOST and that must be reflected in price paid...Kris
Hey, Junko ... I think the whole MOST thing is just a temporary aberration (because the mainstream network biz wasn't interested in a "niche" market back in the '90's). It's going to be completely replaced by Ethernet within a few years ... I can't believe Microchip puts anything more than a temporary value on it. Perhaps the strong relationships with the the auto tier-one suppliers is the real value.
A Book For All Reasons Bernard Cole1 Comment Robert Oshana's recent book "Software Engineering for Embedded Systems (Newnes/Elsevier)," written and edited with Mark Kraeling, is a 'book for all reasons.' At almost 1,200 pages, it ...