[AusNOG] Question on Carrier network design / MPLS

[Steven Crockett]

At a wholesale level most tier 2 (and some Tier 1) providers get their interconnections and "aggregate" in capital cities... any MPLS enablement and switching/routing is done in the core of the network with L2/VLAN delivery of individual circuits.
As an example, even though AAPT have their own POP in Geraldton WA (400+ north of Perth) with their own NBN interconnect AND their own DSLAM, an eLAN/MPLS network between two sites in Geraldton (one on NBN, one on DSL) will go via Perth (logically it may "appear" point to point to the customer, but the latency makes it clear it is not - and they have confirmed that this is in fact the case).
As a Tier 2 - we get little choice on Telstra DSL to use a CBD based LNS, and aggregation circuits due to price and accessibility.

My question is when it comes to Telstra (Retail) - are they any different?  I know they run their big CRS units like most carriers in capital cities (with an added POP in the Pilbara in WA)... but if two customer on fibre/DSL or DSL/DSL or SHDSL/fibre in the same city try to do a point to point connection over an MPLS network the traffic will still travel via the closest aggregation point / capital (Perth)?  Do they still route all DSL traffic via CBD based LNS, or can they deliver a local L2 handoff to MPLS capable units to deliver local routing within an Exchange or zone?

>From a best practice network design it may pay to aggregate and push MPLS closer to the edge, but in practice I don't think the carriers typically do much more than L2 aggregation over MPLS networks with the customer network added in the core??

The customer sees point to point either way - but can Telstra actually deliver lower latency and local routing in these regional areas?

Thanks!

Steve Crockett
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.ausnog.net/pipermail/ausnog/attachments/20150703/b6d27e30/attachment.html>