[72673] in North American Network Operators' Group
Re: Surge Protection
daemon@ATHENA.MIT.EDU (Marshall Eubanks)
Thu Jul 22 16:26:25 2004
In-Reply-To: <6.1.2.0.2.20040722130224.07588ec0@mail.amaranth.net>
Cc: nanog@merit.edu, "Fisher, Shawn" <SFisher@Bresnan.com>
From: Marshall Eubanks <tme@multicasttech.com>
Date: Thu, 22 Jul 2004 16:25:57 -0400
To: Daniel Senie <dts@senie.com>
Errors-To: owner-nanog-outgoing@merit.edu
A good principle is to only let fiber links into your buildings.
This is especially a good idea with roof mounted satellite or
P2P microwave links, which otherwise are basically lightening rods
attached to your routers / rf equipment / whatever. I thought this
was a PITA when I first encountered it with the Navy but it
works and saves a lot of grief.
On Jul 22, 2004, at 1:06 PM, Daniel Senie wrote:
>
> At 11:56 AM 7/22/2004, you wrote:
>
>> Have anyone experienced hardware failure related to electrical spikes
>> coming into your datacenters or equipment locations via the telco
>> facilities? I am referring specifically to copper facilities for
>> DS1's, etc. I know that the telco must maintain good grounding, but
>> sometimes when you get hit with a few Gigavolts worth of electrical
>> energy not much will help you. Whatever the case, has anyone had any
>> experience good or otherwise with surge protection for their Telcom
>> circuits? I am looking at this unit below as a possible solution.
>
> Rule #1, don't trust the telco or the power company, or anyone else
> feeding wires into your building to do a good job keeping you safe
> from surges.
>
> A client of mine has what used to be a CSU/DSU... now has surface
> mount components missing and the like. They hadn't installed a surge
> protector on the T-1. They had covered the power and the antenna
> coaxes at the site. Only the T-1 line was unprotected. Lightning will
> find that one path you've not protected.
>
> The cost of installing a surge protector is unlikely to impact your
> bottom line. One successful lightning strike on the other hand will
> hurt quite a bit, and probably happen at 4AM just to be more annoying.
>
Regards
Marshall Eubanks
