To: signist who wrote (5841 ) 11/9/1997 3:50:00 AM From: Mike Winn Read Replies (1) | Respond to of 42804
May be I should pretend to be Dan Spillane for a moment and come to you guys rescue regarding this GE tech issue. First, we need to define multimode and single mode. Let's imagine an optical fiber like a cyclindrical tube where a light beam emitted from a light emitting diode (LED) or a injection laser diode (ILD) traverses along the tube. Rays are reflected and bounced off the wall of the fiber. Some rays energy is even absorbed by the surrounding material. The result is each ray takes a different propagation path and arrives at the destination at different time, therefore limiting the distance at which data can be accurately received. This is called multimode transmission. If we reduce the fiber core radius, then there will be less angles to reflect. The light beam tends to propagate in a more straight line. This will provide superior performance to multimode transmission and is called single mode transmission. Single mode fiber has core diameter of 2-8 micro-meter (micron) while multimode fiber has core diameter of 50-125 micro-meter (micron). Single mode fiber is more expensive than multimode fiber. Common multimode fiber measures 62.5 micron in diameter while less common fiber measures 50 micron. The issue reported by the GE committee as reported by PC Week is that data can be transmitted reliably only for 100 meters in multimode and this is pretty short distance. There are several fixes that can be worked on such as varying the wavelength of the light source, use better type of light source (laser is better), devise better multimode fiber, or change the protocol. Anyhow, this is a pretty serious issue and cause for concern. It's kind of dumb to discount the PC Week article in this case and call it the National Enquirer. PC Week is simply reporting the news from the GE committee in this case and it's not making up the tech issue. So how does this affect MRV if this turns out to be a real problem? My take is MRV is not affected too seriously in the short term yet. MRV sells the MegaSwitch which can be used with ATM and later on upgradeable to GE. So the user can still buy MegaSwitch and wait for the fix before upgrading to GE. But in the long term, if this problem is not fixed yet, MegaSwitch will loose its attractiveness and MRV will suffer. Btw, the Arberdeen Army testing center mentioned in this article uses the switches from MRV. If I have a big position in this stock, I will surely call Noam and quiz him real hard. In the past when I talked to him, he always stressed the simplicity of MRV's design. Well in all the tests conducted by the tech magazines, MRV's products always fall in the middle or at the bottom of the scale. Price wise, MRV seems to be at the top. So, this is not necessarily bad. May be, MRV's strategy is to target the low end rather than the high end of the market and be competitive in price. But I will surely quiz Noam to see if this strategy would work in the long run. It's bad to fall in love with a stock and be blind sighted. Never under estimate the competition. MRV uses smart engineers from Israel, but so do 3Com, Cisco, Bay, etc. This problem with GE will slow down its deployment and will allow the competition to catch up with MRV. I feel lucky not to be in this stock at the moment. Mike Winn.