[LINK] NBN layer 3 services

[Richard Chirgwin]

Scott Howard wrote:
> On Fri, Dec 11, 2009 at 2:37 AM, David Lochrin <dlochrin at d2.net.au> wrote:
>
>   
>> A G-PON seems like a wise choice because at least the basic network is
>> ~simple~ and therefore reliable.  The deeper the common end-user protocol
>> stack which is layered upon it the more complex it becomes, with potentially
>> greater risks for massive stuffups, not to mention the technology lockin to
>> Cisco.
>>
>>     
>
> And is that simplicity (if it even exists) worth the additional cost - both
> in CapEx and OpEx - for a PON?  How about the distance limitations of PON
> (10-20Km v's 80Km for Active)?  Or the limited diagnostics available with
> PON?  Or the reduction in bandwidth that PON gives? (How many times have we
> heard people complain about how Telstra's Cable is shared bandwidth - just
> like PON is).
>
> Whilst there are some small advantages with PON they ar e normally far
> outweighed by the advantages of an Active Network.
>   
The Comms Alliance reference model for the NBN supports *both* PON and 
direct fibre.

But really, given the stated aim of the NBN is covering 90% of the 
population, a 20km footprint is easily enough. If at some point the 
remaining 10% gets fibre, direct fibre will still be an option (but we 
will never get fibre-to-100%).

As for the "shared bandwidth" problem: it's a non-issue. We all share 
the bandwidth, all the time. That's the nature of the Internet. And we 
haven't even scratched the surface of the theoretical available 
bandwidth in fibre - there's vast room for upgrades.

Returning to MPLS, I can't consider it an immature technology. I suspect 
we're all users of it - in the core of the networks where we don't see 
it happening. Telstra, Optus and AAPT have all deployed MPLS, years ago. 
And Verizon (which carries international traffic). And many others. It 
was immature - years ago, when a company called (I think) IPsilon 
invented the precursor technologies; but that was in another country, 
and besides ...

RC
> Of course there is technically no need for only one technology to be used.
> In many regional areas Active is a clear-cut winner due to the extended
> distances and the better diagnostics available. In metro areas there may be
> value reasons why PON is better (although I can't think of many...)
>
>   Scott.
> _______________________________________________
> Link mailing list
> Link at mailman.anu.edu.au
> http://mailman.anu.edu.au/mailman/listinfo/link
>
>