Muni Fiber and Politics

Scott Helms khelms at zcorum.com
Tue Jul 22 19:23:14 UTC 2014


Mikael,

Let me see if I can clarify for you.

"I don't know where to start. Either you do one vlan per customer and use
very expensive gear that scales this way, or you do several customers per
vlan and do DHCPv4/DHCPv6 inspection (see for instance
http://tools.ietf.org/wg/savi/ documents). Does this answer your question?"

First, QinQ VLAN scaling hasn't been a problem in about a decade nor is it
hard to split out the VLANs to hand them off to other providers.  Second,
all of the gear vendors that I've worked with already have methods for
handling source verification and port isolation if you don't want to do
QinQ.  Certainly any of the "traditional" vendors of broadband gear will
have answers for this already and unless you're planning on grabbing some
enterprise class shelf and jamming it with long range lasers (which most
won't take) you don't have a problem.  Even the Cisco ME line, which is
pretty damn cheap, does this by default

http://www.cisco.com/c/en/us/td/docs/switches/metro/me3400/software/release/12-2_25_seg_seg1/configuration/guide/3400scg/swtrafc.html#wp1038501

"If you're aggregating 10-20k apartments in the same place, I think this
warrants proper space and trained engineers to do the cabling."

The chances that a muni network in North America has both 10-20k apartments
and needs to build its own fiber are pretty much non-existent.  We don't
have the population density that exists in much of Europe and our cities
are much less dense.

"This worked for the PSTN companies, why wouldn't it work for
municipalities?"

The economies of scale are completely different for one thing.  Second, the
phone companies designed their land purchases and buildings around doing
wiring centers and central offices, the cities have never had this need and
most don't have a suitable building (power, cooling, and security) that
isn't already occupied.  That's why its _much_ easier to let the ISPs bring
in some fiber and let them hold all their gear at their site.



Scott Helms
Vice President of Technology
ZCorum
(678) 507-5000
--------------------------------
http://twitter.com/kscotthelms
--------------------------------


On Tue, Jul 22, 2014 at 3:08 PM, Mikael Abrahamsson <swmike at swm.pp.se>
wrote:

> On Tue, 22 Jul 2014, Scott Helms wrote:
>
>  BCP38 (and BCP64) have nothing to do with who is doing layer 2 since
>> neither of those technologies pay any attention to the layer 2 network
>> anyway.  I'd be curious to see your reasoning as to why it needs to be done
>> between layer 2 and layer 3 given that all of the access gear, including
>> the Ethernet equipment, has layer 2 enforcement of layer 3 information like
>> DHCP and static assignments of IP addresses.
>>
>
> I don't know where to start. Either you do one vlan per customer and use
> very expensive gear that scales this way, or you do several customers per
> vlan and do DHCPv4/DHCPv6 inspection (see for instance
> http://tools.ietf.org/wg/savi/ documents). Does this answer your question?
>
>  Keep in mind that in most places a muni network is currently feasible
>> that muni doesn't have a telco quality wiring center in place already and
>> where cities have the resources to build one the market usually doesn't
>> need them to.
>>
>
> If you're aggregating 10-20k apartments in the same place, I think this
> warrants proper space and trained engineers to do the cabling.
>
> This worked for the PSTN companies, why wouldn't it work for
> municipalities?
>
> --
> Mikael Abrahamsson    email: swmike at swm.pp.se
>



More information about the NANOG mailing list