|
Is this optical fiber supported by the DECNIS :
62.5/125 micron multimode optical fiber
numeric aperture 0.275
bandwith 300 MHz/Km at 1,300 nm !!!!!
attenuation 1.5 db/Km at 1,300 nm
Regards
I need answers to this question and the questions of 690.0 before
tuesday (if possible). Thanks.
|
| < Is this optical fiber supported by the DECNIS :
<
< 62.5/125 micron multimode optical fiber
< numeric aperture 0.275
< bandwith 300 MHz/Km at 1,300 nm !!!!!
< attenuation 1.5 db/Km at 1,300 nm
<
The problem with this fiber is, as you realize, the 300 MHz-Km
bandwidth. This is significantly different from the 500 MHz-Km
value required for our FDDI systems. The effect of using a lower
bandwidth optical fiber will be to reduce the length of fiber over
which an acceptable Bit Error Rate can be maintained.
Therefore we can not support use of this optical fiber for a full
2.0 kilometer link.
Chris Baldwin wrote a memo detailing supportable distances for fibers
that do not meet the PMD requirements. It deals with alternate fiber
sizes primarily, but also discusses the lower bandwidth fiber. The
relevant paragraph is reproduced here.
4. Up to 1.6 km lengths of 50,62.5,85/125 or 100/140 fiber that
was not characterized for chromatic dispersion can be
supported provided that the installed cableplant has a modal
bandwidth of 200 MHz at 1300 nm over its full length.
The fiber in question should meet this criteria.
That's the official stance.
Dick Kirk
FDDI Engineering
NAC
LEVERS::KIRK
|
| I'm not sure if it's a good idea to give out details of the internal
architecture such as processors used etc, since they have no way of
knowing how we use them. Anyway here are some answers:
> * which processors are there in the DC (68000 for general purpose + ?)
>and in the DB (68020 for general purpose + ?) ?
DC - 68000. DB - multiple 68020's working in parallel, and a bit-slice
processor to handle full-speed FDDI.
> * MTBF of the DC and the DB ?
For the DB - around 25000 hours. Not sure on the DC.
> * time to be ready (after a power on) for the DC and the DB ?
Around a minute
> * how many adresses are supported on the DB ?
~14K
> * are all the protocols (DECNET, IP, OSI, IPX, SPX, ATK ...) translated
>by the DB ?
Yes
> * are the configuration of the DC and DB (IP, trap ...) and the adress
>and protocol filterings for the DB kept in the DB (without need to load any
>thing at power on) ?
Yes. If the IP address is administrated via Bootp, the device will
get its address from a server every time it comes up. (but its kept in
non-volatile RAM if configured through ELMS)
> * is there any limitation to configure and manage DC and DB only with
>SNMP from an HP OPENVIEW station ?
Can be managed by any NMS that complies to SNMP
> * is trap mechanism only available with SNMP protocol (no "trap"
>with RBMS or Out Of Band Management)?
The ELMS poller provides functionality that simulates alarms by
polling critical variables at regular intervals. There's no
asynchronous notification from OBM.
> * what info do we have with leds on the DC and the DB (sorry, i have no
>documentation here)?
From memory, there's a good/bad LED per FRU, a working/not-working
LED per FDDI port, a forwarding/broken LED for every bridge port.
There may be others I've missed.
Anil
|