Is there more than one standard for passing 100 Base ethernet over fiber? I have a media converter that says it will convert from 100Base-T to
100Base-FX at 850nm. The device I need to plug this to (a DWDM system) has a tributary that supports FX-100. It's not clear to me if these are identical, or if there might be more than one standard for support of fiber optic 100 Megabit ethernet.
Yes, there is -SX, -LX, -SZ; some add -LH as distinct from -LX (but they are often implemented together.) Then there are the wave division varieties.
SX is usually multimode only. LX ... I haven't looked it up for awhile, but I seem to recall LX can be over multimode or single mode. SZ is single mode.
There are different fibre sizes permitted, and different fibre qualities involved, and the frequencies can differ, and the combination of frequency, quality, and fibre size determines the maximum distance.
There is only one "standard", but lots of non-interoperable 1.2 Mbaud transceivers out there.
For Gigabit, there are a plethora of standards (SX, LX, LH, ZX) and plenty of custom implementations (e.g Cisco's CWDM GBICs, or Champion's extra-long reach GBICs). But, as far as I know, the only official standard for 100mb is 1310nm coded as 4B/5B on a 1.2MBAUD carrier.
Back in 10BaseF days, there were three standards plus one pre-standard: FOIRL was pre-standard (championed by Digital), 10BaseFL was the mainstream varient (850nm, Manchester encoded), 10BaseFB was widely available (both Synoptics and ... Sorry, can't think of the big 10BaseFB pusher). I think it was also 850nm, but it had some sort of syncronous clock instead of Manchester. Then there was Codenal with their truly bizarre 10BaseFP variant. It used much "hotter" LED's and passive hubs which were made of fiber with the insulation stripped off and wound tightly together. The idea was a half-duplex passive star where all of the stations would see the signal bleeding across pairs.
That merely describes the fiber, and says that it is ideal for 1310nm communications (like FDDI, 100BaseFX, and 1000BaseLX) but also works fine for 850nm applications (such as FOIRL, 10BaseFL, 10BaseFB, and 1000BaseSX)
It does not describe a PHY that can be used to transmit 120MBAUD at 850nm.
According to the FOLS whitepaper, "100Base-SX Fast Ethernet White Paper: A Cost Effective Migration Path for Fiber in the Horizontal" "The signal encoding is identical" (to 100Base-FX)
The fiber type for 100BASE-FX *is* specified in IEEE 802.3; it is precisely the same as that in FDDI, as the FDDI PMD is incorporated into the IEEE standard.
"The 100BASE-FX PMD (and MDI) is specified by incorporating the FDDI PMD standard, ISO/IEC 9314-3: 1990, by reference ... ." IEEE 802.3-2002 ? 26.2.
-- Rich Seifert Networks and Communications Consulting 21885 Bear Creek Way (408) 395-5700 Los Gatos, CA 95033 (408) 228-0803 FAX
Thanks for that statement. And it would be interesting to see this ISO/ IEC standard (...but unfortunately more difficult to have because IEEE
802.3 is free).
Anyway, do you agree that 100BASE-CX is not a good IEEE namming?
Could I take the opportunity to ask you something else: one day, and perhaps there is a long time ago, you wrote this:
"...In 100BASE-T, the line signaling is synchronous; there is an active-idle signal on the line at all times, maintaining receiver clock synchronization even in the absence of data frames. This is, in part, why 100BASE-T cannot support shared-media wiring systems"
I will be happy if you could explain me (and us, of course...) just the last point "cannot support shared-media wiring systems". Thanks very much for your words.
Yes, the preamble is there, even for gigabit and ten-gigabit ethernet. I recall Rich posting on that point in the early days of gigabit, when I was arguing with him that Cisco's 1000BaseCX product, "The gigastack gbic", constituted a bone-fide commercial half-duplex gigabit implementation. Thankfully, I haven't had to deploy one of those in years.
Yes, I like the naming system more like 100baseCX. (snip)
For 10base5, as an example, more than two stations attach to a single coaxial cable. (Well, more than one is enough to qualify as shared.) For 10base? standards, there is no signal on the cable when no hosts are transmitting.
I don't see anything in that statement that would prohibit a non-contiguous signaling system for coaxial 100 megabit ethernet.
Is there still a preamble for 100base??, as would be needed to lock the PLL onto the signal?
There is, but it is not needed for acquiring phase or frequency lock, since continuous signaling allows the receiver to remain locked over arbitrarily long periods of time. The preamble was left in the frame format for historical and pedantic reasons.
-- Rich Seifert Networks and Communications Consulting 21885 Bear Creek Way (408) 395-5700 Los Gatos, CA 95033 (408) 228-0803 FAX
If one ever did develop a shared media system, though, then it might be useful. I suppose that is less and less likely to happen, though, as UTP systems get cheaper.
No, it is true for 10BASE-T as well. While there are link pulses, these occur for only 100 ns out of 62.5 ms (line is idle 99.9998% of the time), and even the link pulses are not transmitted if there is sufficient data intensity. In any case, link pulses are not repeated by a hub, i.e., they are local to the dedicated twisted pair medium.
-- Rich Seifert Networks and Communications Consulting 21885 Bear Creek Way (408) 395-5700 Los Gatos, CA 95033 (408) 228-0803 FAX
Cabling-Design.com Forums website is not affiliated with any of the manufacturers or service providers discussed here.
All logos and trade names are the property of their respective owners.