Why all the JINI questions?
First, I very much appreciate the help and direction from Charles and Eric. I've not found my answers yet, but I will in time.
Maybe I'm looking at things a bit differently. One thing you learn after 25 years in semis is that everything is do'able at a price. Much of what you see in consumer land today was possible five years ago, but at too high a price. Yes, the esoteric stuff is available at the cutting edge, but I'm talking about mainstream consumers; the the brass ring! This is why I think JINI questions are important.
You hear talk about cars having over 20 "microprocessors" and houses having potentially hundreds. They even tell us about those in shoes and clothing. They are here and they are now, but they aren't really microprocessors; they are micro controllers.
Micro controllers are typically very limited function 4 and 8 bit controllers with small masked ROM (Read Only Memory), a little bit of scratch pad RAM and maybe some clock, driver, etc. functions. They don't run an O/S like Windows or any other you might know. Many run compiled machine code to minimize the need for memory. These parts most often sell for between $0.50 and $1.50. They are what you find in cars, washers, stereos, exercise equipment, cordless phones (they use two), etc. From what I've learned thus far, I seriously doubt these simple controllers will support the JINI vision.
For a product to run JINI directly it will need to utilize a more powerful processor and will probably also need to be equipped with more memory. This adds overhead and cost directly to each device. Combining this with development costs, lack of a ready market and rising interest rates makes for a daunting and most likely slow ramp to volume production. IMO, it becomes obvious we won't be seeing the world switch to JINI toasters any time soon. However, I expect (hope) some high end stuff will become more common.
JINI is, IMO, the winning concept. JINI allows a device to identify and log itself onto a ad hoc (community) network and share resources with other members. It allows devices to control other devices without possessing drivers; almost as though it was instigating teamwork between inanimate objects. It adapts to changes, is intended to work with any O/S and can alert operators to problems.
JINI is the logical extension of client side JAVA (actually just now starting to take hold) and a dream for the future that should be quantified. In this case, I feel determining its cost of implementation is one of the critical steps we need to take when estimating it's present value and building a model for its future. By knowing its overhead cost in hardware and development, we can better judge how quickly and at what level it will be implemented.
Regards, pmcw |
