Post-Event Report from the Embedded Systems Conference 1999

Linux Weekly News
LWN daily updates
Linux Stocks Page
Contact us

Post-event Report from the ESC West 1999

San Jose, CA USA October 4th, 1999
Let us start with errors or potentially misleading information that was included in our earlier reports. First, my estimate of attendance at the conference was grossly inaccurate. Although final figures for the event are not yet available, the most likely number is between 14,000 to 15,000 warm bodies. That means a number that includes all attendees, exhibitors, press, people running the hot dog stands, etc. These numbers came from Scott Minor at MFI (the company that runs the show), so they should be reasonably accurate. My apologies for being off by a factor of four! However, I stand by my comment that this show was two to three times the size of LinuxWorld, at least in two categories: the number of exhibitors and the number of classes offered.
Second, my coverage of Cygnus' plans may have given a false impression. Although they do not plan to come out with their own Linux distribution, they plan to "support the embedded Linux platform in such a way that _any_ Linux distribution can be embeddable," according to Michael Tiemann. They will be developing their own Software Development Kit (SDK) and configuration technology, so they do, in fact, expect to be in direct competition with Lineo in this area. Michael commented, "We believe we have the advantage in that we've successfully created open source configuration technology in the past: GNU configure, GNU autoconf, and non-open source graphical eCos configuration tool (the text version is open source)".
I got an opportunity to talk to several different engineers after the show. These weren't people currently involved in Linux. These were engineers that develop in the embedded systems market now, using proprietary operating systems and development tools. All of them thought Linux was a "cool thing"; all of them would prefer to be developing for Linux rather than for NT; most of them were already using Linux at home. However, these are engineers that are accustomed to having slick, GUI-based tools for their work. They wanted to know what Interactive Development Environments (IDEs) existed for Linux and for what compilers. Many of them were very attached to their GUI-based tools and not interested in command-line versions. It was clear, based on this, why both Lineo and Cygnus are looking at the tool market. There is also a lot of money involved! Prices quoted to me ranged from "$22,000 up front, plus a $3,000 annual fee" to "$10,000 per year" for various support packages from vendors. Given both the potential money and the level of interest, it is easy to see why both Lineo and Cygnus are looking at the tool market.
My impression is that, in many ways, embedded Linux is still in its infancy, just as the Linux desktop is just beginning to take off. Linux will do extremely well right now in very specific cases, but does not yet have the tools needed to be of major concern to the current players. The embedded systems market that I saw at the conference was highly fragmented, but full of vendors that are accustomed to heavy competition. They provide full-featured, mature tools and good support in order to distinguish themselves from their competitors. This is why the embedded Linux market is in such an embryo form; the proper tools will need to be developed before it can actually challenge existing accounts just as desktop Linux needs office productivity tools to rival the Microsoft Office suite before it can be an option for everyone to consider.
That said, the open source model guarantees that we will get there, since the value that is put into open source tools will not be lost. Right now, if any of these proprietary embedded systems tool vendors goes out of business, they may take all the work of their developers and put it into a black hole. That won't happen with the work that is contributed in an open source form, so we will eventually catch up with, and surpass, the proprietary model. No timeframe for that happening is guaranteed, of course, but vendors like Cygnus, Montavista and Lineo are well-poised to make it happen sooner, rather than later. Cygnus has a proven track record as an open source company for which it would be difficult to find an existing match. Montavista has made clear their commitment to and belief in the open source model as a critical element of their business. I challenge all three of them to keep what they develop open source, in recognition that their current and future prosperity will depend on it.
--Liz Coolbaugh