Editorial 1Q97
BUS WARS AGAIN!
A new look, as you can see. Indeed, as the magazine starts its 11th year of publication, we should do something to mark this anniversary. Today, Real-Time Consult is responsible for 2 publications: this magazine and the Mezzanines Magazine supporting the GRoupIPC organization.
Distributing information in an independent way has always been our objective.
To give the reader an overview of what today's technology has to offer and to help them to make sound choices for the future. This magazine intends to give an overview of current bus technology. And we are just in time. Literally a WAR is going on today between VITA and PICMG, between VME and CompactPCI (CPCI). It's again the old story about bus bandwidth. VITA (VMEbus) is now claiming a 320MB/s breakthrough and is therefore doing better than CPCI is able to do today.
But is this the only issue? It is fundamental to first recognize what architecture you need.
Therefore, I have written this article about the right bus in the right place. When discussing on the one hand profiles 1 & 2 (or processor bus, with or without cache coherence support), bandwidth is an important issue. Looking at multiprocessor applications (profile 3 & 4) on the other hand, bandwidth is not so much of an issue.
The first question is not whether you need 320MB/s but what System architecture you want to develop. If you go for a simple architecture for your System, only a few boards are involved and PC-like software is an issue, then PCI and related architectures should be considered. Don't forget that PMC as a mezzanine PCI solution is a very modular and effective way of solving this problem today without going in the PCI backplane discussion!
If you go for a complex architecture, then VMEbus is and will be for a long time the only solution. This combined with PMC gives the best of both worlds! Both 320MB/s VMEbus and high speed CPCI have still to demonstrate some stability of technology. The least we can say about this WAR is that it shows that this stability has not been achieved. So don't construct your future on quicksand.
This WAR was launched at the Real-Time Computer Conference in Santa Clara. Simultaneously there was a Real-Time Systems Conference in Paris. The trend of low attendance over the past few years at the Paris show was re-confirmed this year, and in Santa Clara the expected number of participants was certainly not reached.
Although more than one reason can be advanced for this, let us concentrate on just one. Engineers are over-worked and have no time left to attend this kind of show and conference. So what should you do instead? You use the WEB and read good magazines. No single company in this field can afford NOT to have a WEB site promoting its products. VITA's Virtual Trade Show, our Real-Time Encyclopaedia and other "Internet"-based initiatives are definitely helping to lower the attendance at these international events.
That is the reason why we have decided to put more effort into our Real-Time Encyclopaedia which intends to be an on-line complement to the magazine.
This year we will also launch product evaluations and discussions. In the next issue, which a is Real-Time OS one, we will go into detail about the hot discussion about the use of Windows-NT in a real-time environment. If you cannot wait, go and see the Real-Time Encyclopaedia.