To: White Shoes who wrote (313 ) 1/9/1998 12:58:00 AM From: BenYeung Respond to of 598
White Shoes, my list of 151-200 wont due until this weekend...alot of DD to do while I have limited time. However, I have done more research on ATM, from two professionals and from 2 books, both sources have shown a bright future for this infant product. Gigabyte Ethernet will dominate for a while due to low price and can use current switches, routes and hubs. But GE will still experience congestion (eg. when you download a file, U get say, 30% in the first second, then you have to wait for another 20%, then 35%...because the datas are chopped into unequal pieces to be transferred in different routes to the same destination) due to the limitations of frame relay. In a book on frame relay, it dedicated a whole CHAPTER to explain how to deal with congestions. It shows how slow this 13 year technology is. Cell relay in ATM are continuous flow of equal-sized cells (when there are no data, it transfers idle cells). Speed sensitive cells (voice) are tranferred first. But if the network experiences congestion, the system will selectively toss out cells of the speed senitive data to insure timely delivery (Human ears cannot notice the loss of a few cells). The quality sensitive cells (data), however, are the last thing (probably never) the congested system will toss out to insure quality transfer of data. This alone solves the network congestion really well. Since ATM works the best in high quality/high bandwidth medium such as fiber optics (its continuous cell flow takes quiet a bit of bandwidth), Yurie Systems has a patent to enhance ATMs' speed in copper/old-style cables. Look, I am not promoting YURI because I own none of it (but I will), but I am tell you all to watch out for this new tech (actually it has been around for a while, developed by IBM and Bell). Invest in companies that is/will become a player in this new tech. One author describe the Information Age as BATM (Before ATM) and AATM (After ATM). ATM own only 1-3% of all networking in 1994 while Ethernet took 90%+, but the author is anticipating a complete reverse.