x
<<   <   Page 4 / 6   >   >>
jim_smith 12/5/2012 | 1:21:01 AM
re: Stitt: Sonet's Even More Dead Ethernet is getting better but it has a long ways to go as well.

There is no fundamental technology difference between XYZ/MPLS/Ethernet vs. XYZ/SONET in terms of providing service to the end user. It is only an engineering issue.

Cisco is trying very hard. If they keep on trying, they will get there eventually.

I don't think it will happen within a year, but I don't think it will take 10 years either.

This might be one of those rare occasions where Europe/Japan/South Korea end up leading the way.
somedumbPM 12/5/2012 | 1:21:01 AM
re: Stitt: Sonet's Even More Dead You netheads be careful with those GBICs that go 4000km without regen. You could poke an eye out - hehe.

Also the FCC fines telcos by a formula for outages. Something like # of lines down X length of time down X an amount that will make you say 5 9's isn;t enuff and if it's E911 office yikes! That'll make the water pretty cold when ya first dip ya toe in it.

Nothing like having a virus take out a call server at a 911 call center and no analog lines around.

Sonet may die, but it will not be anytime soon. It will be interesting how it all plays out for sure though. Ethernet is getting better but it has a long ways to go as well.
jim_smith 12/5/2012 | 1:21:01 AM
re: Stitt: Sonet's Even More Dead Ethernet and IP services are certainly making inroads in replacing T1s, T3s, and OC-Ns as a service. The vast majority of those services, however, still relay on SONET to move the bits accross the city or the country. This is not to say that some form of Ethernet (or something else) will not replace SONET some day, but it has a long way to go.

Let's not get too carried away with this kind of thinking.

MPLS/Ethernet may prove to be SONET's undoing. MPLS/Ethernet has Cisco's backing, and I don't think it is a good idea to underestimate Cisco.
signmeup 12/5/2012 | 1:21:00 AM
re: Stitt: Sonet's Even More Dead jim_smith,

Unfortunately, engineering issues usually dictate service capabilities. How do I put an ethernet port in remote loopback?

Until someone defines a way make management ubiquitous across the last mile, ethernet still has a long way to go.

signmeup
douggreen 12/5/2012 | 1:20:55 AM
re: Stitt: Sonet's Even More Dead Jim_smith,

The success of MPLS does not neccesarily mean the undoing of SONET. Both IP and MPLS can run over either Ethernet or SONET as a transport layer.

I never underestimate Cisco. However, Cisco has yet to prove that they can translate their dominance in datacom to a dominance in Telecdom. I wouldn't count them out, but they cannot dictate what happens in Telecom like they can in datacom.
douggreen 12/5/2012 | 1:20:54 AM
re: Stitt: Sonet's Even More Dead jim_smith,

When I said that Ethernet has a long way to go, I wasn't speaking of the capabilities of the technology. I was speaking of the progress that it has made so far in replacing SONET as an carrier infrastructure.

Of course, both SONET and Ethernet will be replaced some day ;)

By the way, what is your background?
douggreen 12/5/2012 | 1:20:53 AM
re: Stitt: Sonet's Even More Dead Priam,

You make some good points. It certainly makes no sense, in theory, to impose a ring technology on a network that is a natural mesh, such as that in the long distance network. That's what the optical cross connect guys have been pushing for years.

I have no personal interest in seeing SONET live forever. In fact, I spent years trying to replace parts of the SONET network with DWDM. We succeeded in taking a pretty good bite.


priam 12/5/2012 | 1:20:53 AM
re: Stitt: Sonet's Even More Dead Though "SONET is dead" is intentionally inflammatory, the point is that SONET is becoming more and more a link technology, - the rings, so to speak, are becoming binodal. The network is more and more packet oriented, with switches and routers at the interstices, pushing the fat, dumb pipes toward fat, dumb clouds with smart stuff everywhere you can reach...

----->
The success of MPLS does not neccesarily mean the undoing of SONET. Both IP and MPLS can run over either Ethernet or SONET as a transport layer.
priam 12/5/2012 | 1:20:50 AM
re: Stitt: Sonet's Even More Dead But here we are, exchanging a few bytes very transiently. The point-to-point high bw service will always be important; but it's probably more tractable building p2p out of a cloud model than the other way 'round. Hence the importance of MPLS/pseudowire. Over the last ten years or so, the really interesting stuff has been falling out of reachability, not (just) speed.
-------------->
If I put a packet into a fat pipe, truck it all the way across the country, and then find out that it doesn't fit into the smaller pipe through which it's being delivered, I am wasting money. And if I've dropped another viable packet at the source end in order to transport that waste, I'm not making my customers any happier.

If I need to re-examine those packets a couple dozen times enroute (with very low entropy) because I've replaced all my SONET regens (and muxes) with routers, I'm not exactly using resources effectively either.

PO 12/5/2012 | 1:20:50 AM
re: Stitt: Sonet's Even More Dead Sure it's fun to ignore what networks have been carrying for the last several decades, but to ignore what networks have to deliver going forward seems a bit shortsighted.

If I put a packet into a fat pipe, truck it all the way across the country, and then find out that it doesn't fit into the smaller pipe through which it's being delivered, I am wasting money. And if I've dropped another viable packet at the source end in order to transport that waste, I'm not making my customers any happier.

If I need to re-examine those packets a couple dozen times enroute (with very low entropy) because I've replaced all my SONET regens (and muxes) with routers, I'm not exactly using resources effectively either.

As more traffic shifts to UDP, there is an increased need for traffic engineering on the network, and the more times you open the pipe the more complex that problem becomes.

The value of multiplexing has not been diminished. The only question is whether the cost of installing and managing multiple layers of multiplexing pays off. And I see no reason to believe that managing a SONET interface on a router will become simpler and cheaper than managing the same interface on a SONET mux.

The better headline would come from the Monty Python sketch ... Not dead yet? Getting better!
<<   <   Page 4 / 6   >   >>
HOME
Sign In
SEARCH
CLOSE
MORE
CLOSE