Cisco, Sycamore Circling Lucent's ATM

COMMENTS

Newest First | Oldest First | Threaded View

<< < Page 7 / 7

skeptic 12/4/2012 | 9:27:21 PM

re: Cisco, Sycamore Circling Lucent's ATM On the technical side, can MPLS really provide the fine-grained QoS that ATM provides?
What about the OAM&P features currently available through ATM?
-----------------------------
If you talk to the people who buy MPLS today,
none of them seem to want the OAM&P features.
They are buying it for other reasons.

If you talk to ATM people trying to convince
them to buy MPLS, its mostly not productive
because they don't want anything but ATM.
Any ATM feature that MPLS doesn't do is a reason
not to use MPLS.

As far as fine-grained QOS, there is no technical
reason why MPLS can't provide it. (aside from
the QOS issues that arise from allowing larger
packet sizes). Its a connection-oriented signaled
protocol.

Reply | Post Message | MESSAGES LIST | START A BOARD

skeptic 12/4/2012 | 9:27:19 PM

re: Cisco, Sycamore Circling Lucent's ATM Take Frame Relay and brush liberally with NIH. Sprinkle with gratuitous jargon designed to make the same things sound different.
----------------
Or....take frame relay and fix it. Update the
standards to something reasonable and remove
all the "junk" that was bundled with FR but
of no use to many people.

As far as gratuitous jargon, the FR standards
are far WORSE than any of the IETF documents
for MPLS. They are not written for humans,
needlessly duplicate each other (did we really
need three types of LMI?) and are full of
features that nobody uses/wanted.

If the people using FR and the standards bodies
associated with FR had been interested in
improving it, MPLS might not have happened. But
(as with the ATM community) they were not
interested in doing anything or changing anything.

Reply | Post Message | MESSAGES LIST | START A BOARD

indsavvy 12/4/2012 | 9:27:17 PM

re: Cisco, Sycamore Circling Lucent's ATM one thing misssing from both previous posts. FR is proven and revenue generating---majorly revenue generating. MPLS is still a wannabe, just ask all the MPLS switching product vendors who are now touting ATM on their boxes.

Reply | Post Message | MESSAGES LIST | START A BOARD

zoinks! 12/4/2012 | 9:27:15 PM

re: Cisco, Sycamore Circling Lucent's ATM >Sycamore would get an instant revenue stream
>which they are not getting right now from their
>current portfolio.
>blah blah blah

Plus, many of the SCMRys jumped ship from Ascend to Sycamore and are well-versed in both the GX/CBX/BSTDX and NavisCore architectures and code.

...now whether they would want to put their hands back into the spaghetti is another story...

Zoinks!

Reply | Post Message | MESSAGES LIST | START A BOARD

putridrage2000 12/4/2012 | 9:27:04 PM

re: Cisco, Sycamore Circling Lucent's ATM MPLS will only be of use if telcos can efficiently provision services on it that bring to their end users costs savings. If I have a layer 2 Frame/ATM network at a recurring monthly charge of $100, and MPLS cannot bring to me services that are better, or a price that is better, than what I already have, than what good is it to those who actually pay the bills? Remember all the VPN hype? Sure VPN is great for sitting at home on your cable modem and connecting to the office, but besides that, its still pretty expensive and nobody is deploying it. It sure has some niches in the International arena, but for the most part, Layer 2 networks are still the leader.

Yes, ATM as a growth industry is dead, nobody is rolling out a national ATM network these days. But I am sure more and more end users (once again, those folks that have money and pay the bills) as they migrate away from 56k frame relay or T1 frame relay need to go to some technology. Does anyone have any idea how many 56k frame relay circuits still exist? T1 Frame relay? Thats alot of cards that LU can sell into the embedded carrier environment.

I think MPLS will be too difficult to provision for carriers to offer it at a cheap price. But we will see.

Reply | Post Message | MESSAGES LIST | START A BOARD

brahmos 12/4/2012 | 9:27:03 PM

re: Cisco, Sycamore Circling Lucent's ATM http://digitalmass.boston.com/...

Reply | Post Message | MESSAGES LIST | START A BOARD

sntwk 12/4/2012 | 9:27:00 PM

re: Cisco, Sycamore Circling Lucent's ATM Both have good points and the problem is how to marry them so that carriers who need QoS or support existing customer solutions with minimum impact. As some one pointed out FR is transported on ATM backbone networks. It is unlikely FR will go away and hence it is unlikely that a FR CPE will be plugged away from ATM and connected to a MPLS switch any time soon.

The bad thing about ATM is the in-band header wasting bandwidth. Imagine carrying an IP packet - ATM header, EtherHeader, IP header - customer traffic. So many headers! Consider MPLS - just one or two labels + IP packet. Consider the same for L2VPNs - MPLS Label(s) + FR traffic , plus managing them is and will be easy.

On the other hand consider providing deterministic QoS (latency and peak/avg rates) - it will be challenge to get MPLS switches to deploy such that it happens. I am sure there are lot more advances that happened in the last 10 years or so that should help MPLS switches provide same or better QoS than an ATM switch can. But it is more a risk reward equation rather than technical issue. Do carriers have the patiences and bandwidth to take on a technology that is developing and taking it through deployment, hardening, maturing process that can be anywhere from 3 to 7 years. An MPLS backbone network can help carriers in the longer run reduce costs both operational and capital investment costs as MPLS equipment will be cheaper and configuring and operating them will be easier.

Both ATM and MPLS can't offer fast restoration by themselves. The underlying SONET/SDH or RPR infrastructure must be designed for sub 50 ms restoration. One can't depend on PNNI or OSPF-TE to help fast restoration. PNNI may have an edge over OSPF-TE as it has hierarchical apporach compared to OSPF. The hierarchical approach is what is used in voice networks and their signaling protocol - the #7. Voice networks can setup calls in few seconds and support thousands of calls per minute and millions of calls a day without an issue. It is not the smartness in the signaling systems - they don't depend on routing. There is no routing in signaling #7. They just pre-design bandwidth (the link sets) for signaling transport and the messages go over a deterministic path. Adding some flexibility to signaling #7 (which is what the B-ISDN folks did or try to do) and using that in an MPLS environment with the switches just providing label switching would be more useful. The problem is inability to seperate label switching from control protocols. They must be treated seperately if only to mimic voice switching + #7.

Reply | Post Message | MESSAGES LIST | START A BOARD

broadbandboy 12/4/2012 | 9:26:40 PM

re: Cisco, Sycamore Circling Lucent's ATM Verizon

BBboy

Reply | Post Message | MESSAGES LIST | START A BOARD

network 12/4/2012 | 9:26:38 PM

re: Cisco, Sycamore Circling Lucent's ATM There always has to be a next thing. ATM was the next thing once. Yet even today I haven't seen a widely available ATM svc service from any carrier, worldwide. In fact most commercially available pvc's have little flexibility compared to what the ATM spec's offer.

MPLS just like ATM will emerge in a very simple form, as carriers figure out how to deploy and make money off of the service. All of the bells and whistles of MPLS will be minimized when they hit the end customers.

MPLS good or bad?
MPLS Inevitable. But not near as exciting as all of the hype, GMPLS included.

As for Lucent, keep the business. The investment is over. It's like selling your 10 year old car for $100 when it can carry you to the bank for another ten years. You might temporarily feel better with $100 in your wallet, but you financial statement looks a little better 5 years down the road.

Reply | Post Message | MESSAGES LIST | START A BOARD

billyjoebob 12/4/2012 | 9:26:26 PM

re: Cisco, Sycamore Circling Lucent's ATM Without further developement of the CBX and GX there is no assurance that any buyer of the Lucent ATM gear will have an inside track to update it - or even that Verizon will expand it. It is unlikely that Verizon will replace their 500s and 550s. It also doesn't assure the buyer they will have the inside track on an MPLS core or overlay network.

The buyer will assume the responsibility of supporting these switches for the next 3 to 5 years on what I can only assume will be very thin support agreements. This is a "2 in the bush" proposition.

This would be a good idea - if Verizon wereactually going to buy anything - but they aren't doing any major network upgrades or changes till they pay their own bills down first.

Reply | Post Message | MESSAGES LIST | START A BOARD

<< < Page 7 / 7