|
|
|
|
|
|
|
|
|
|
|
|
| home | help | back | first | fref | pref | prev | next | nref | lref | last | post |
From: Frank Coluccio <frank@dticonsulting.com> To: Rick Kunkel <kunkel@w-link.net>, John L Lee <johnllee@mindspring.com> Reply-To: frank@dticonsulting.com Date: Sat, 09 Sep 2006 00:06:43 -0500 Cc: nanog@merit.edu Errors-To: owner-nanog@merit.edu I would add to the recommendations already stated by Mark and John the folllowing: Depending on the dimensions of the colo in question and the len= gth of the cabling runs you plan to install, take note of the distance limitations= of STP for DS1 signals and the in-house coaxial connections for DS3 signals and higher, which are (were) governed by ANSI/Bellcore, the last time I looked. Fiber, likewise, must be coordinated with network element interfaces (GBICs= , ST, etc.) and here again distance is a major consideration. The latter may beco= me an issue in large structures, or in the event that you plan to run media betwe= en buildings, or if you plan to interconnect with service providers at the f-o= level (Layer1). For your optical media and connection hardware (patch panels, raceways, etc= .), have a long hard talk with at least two reputable fiber optic cable supplie= rs (e.g., Corning, CommScope, Sumitomo, etc.) and develop an understanding abo= ut the limitations and advantages of the various s-m and m-m options you have avai= lable to you, per the types of solutions you need to implement and the distances = they dictate. Incoming fiber from the street (dark fiber providers, included) wi= ll be single-mode, primarily, but the preponderance of your in-house cabling betw= een switches, routers and servers, if fiber distances are indicated, will be multi-mode, requiring a different patch bay selection.=20 Where distances permit (<100 meter channels) UTP should suffice for Ethernet speeds up to 10Gb/sec now, with 10Gb/sec (10GBASE-T) ratified only recently= (I'm quite certain), but I suggest reading the following article from Cabling Installation & Maintenance Magazine, just the same.=20 http://tinyurl.com/mrack Note, Cat5e and vanilla Cat6 will not suffice (perhaps a Cat6A will, not su= re at this stage, but do your homework before purchasing anyone's Cat7) for 10Gb= E (it may work, but your hardware vendors will not honor warranties when problems arise), so be prepared to make some long range planning decisions in coming= to terms with a cabling plant that's going to last you a while. HTH. Frank A. Coluccio DTI Consulting Inc. 212-587-8150 Office 347-526-6788 Mobile On Fri Sep 8 23:31 , John L Lee sent: > > > >=20=20 >=20=20 > > >Rick, > > > >The organization and standards you are looking for are: > > > >BICSI - http://www.bicsi.org/ and TIA/EIA 568 et al for structured >cabling design for low voltage distribution. > > > >The BICSI organization has training and certification for RCDD >Registered Communications Distribution Designer > > > >A BICSI article that is on there web site about data center design is >http://www.bicsi.org/Content/Files/PDF/link2006/Kacperski.pdf. > > > >TIA/EIA 568(ab) how ever many they are up to discuss structured cabling >design for UTP/STP/fiber/coax including patch cables single and multi >pair UTP/STP/fiber patch panels, HVAC control, fire system control >and security systems. > > > >John (ISDN) Lee=20 > > > > > > > >Rick Kunkel wrote: > > > Heya folks, > >I hope this is on-topic. I read the charter, and it falls somewhere along >the fuzzy border I think... > >Can anyone tell me the standard way to deal with patch panels, racks, and >switches in a data center used for colocation? I've a sneaking suspicion >that we're doing it in a fairly non-scalable way. (I am not responsible >for the current method, and I think I'm glad to say that.) Strangely >enough, I can find like NO resources on this. I've spent the better part >of two hours looking. > >Right now, we have a rack filled with nothing but patch panels. We have >some switches in another rack, and colocation customers scattered around >other racks. When a new customer comes in, we run a long wire from their >computer(s) and/or other device(s) to the patch panel. Then, from the >appropriate block connectors on the back of the panel, we run another wire >that terminates in a RJ-45 to plug into the switch. > >Sounds bonkers I think, doesn't it? > >My thoughts go like this: We put a patch panel in each rack. Each of >these patch panels is permanently (more or less) wired to a patch panel in >our main patch cabinet. So, essentially what you've got is a main patch >cabinet with a patch panel that corresponds to a patch panel in each other >cabinet. Making connection is cinchy and only requires 3-6 foot >off-the-shelf cables. > >Does that sound more correct? > >I talked to someone else in the office here, and they believe that they've >seen it done with a switch in each cabinet, although they couldn't >remember is there was a patch panel as well. If you're running 802.1q >trunks between a bunch of switches (no patch-panels needed), I can see >that working too, I suppose. > >Any standards? Best practices? Suggestions? Resources, in the form of >books, web pages, RFCs, or white papers? > >Thanks! > >Rick Kunkel > > > >=20=20 > > >
|
|
|
|
|
|
|
|
|
|
|
|
| home | help | back | first | fref | pref | prev | next | nref | lref | last | post |