Optical/IP Networks

Alcatel Sticks to Core Plans

Alcatel SA (NYSE: ALA; Paris: CGEP:PA) has faced some questions about its core-networking strategy since halting development on its 7770 Optical Broadband Exchange core router. But company officials say they're happy pushing the former TiMetra Networks as a core IP play.

"On the carrier side, we're comfortable with our roadmap," says Basil Alwan, president of Alcatel's IP division and former CEO of TiMetra.

Alcatel froze the 7770 last year and has since said the 7750 Service Router series developed by TiMetra will be its core-routing play, in addition to being a multiservice edge router (see Alcatel Redraws Router Strategy). That leaves Alcatel without a multichassis router along the lines of the Avici Systems Inc. (Nasdaq: AVCI; Frankfurt: BVC7) TSR or the Cisco Systems Inc. (Nasdaq: CSCO) CRS-1. The vacancy is emphasized by the fact that at least one core play was on the block recently: Procket Networks Inc., which recently sold assets to Cisco (see Cisco to Pay $89M for Procket Assets).

But then, a lot depends on what you mean by "core router." The 7750 doesn't pack a terabit punch, but it can stuff 60 OC192 ports into one 7-foot rack (that is, 20 ports per 7750 SR-12, and three SR-12s per rack). That puts it on a par with the Juniper Networks Inc. (Nasdaq: JNPR) T640, which boasts 64 OC192s per rack, Alwan notes.

That's plenty for many customers' core needs, Alwan says. "If we go to a customer that's not one of those 'super' cores, the 7750 is every bit a core router," he says.

Still, doesn't Alcatel hear the call of that "super core" market? Alwan doesn't dismiss the possibility, but he suggests that it's not an imperative for Alcatel: "I don't feel huge pressure just because people are asking about the 'whole solution,' " Alwan says.

Alcatel might have its hands full with the 7750 anyway. The company's announced plan is to move some of the 7770 features to the 7750, but that could prove tricky, according to Geoff Bennett, chief technologist for Heavy Reading.

One problem is that the 7770 was being built from scratch to accommodate high availability, including a feature called the Alcatel Carrier Environment Internet System (ACEIS). It prevents disruption among a router's neighbors even if the entire control plane fails (see Alcatel Bids for IP Core). That kind of functionality "cannot be bolted onto the 7750 -- it has to be designed into the core of the product," Bennett writes in a Heavy Reading report titled, "Ethernet over IP/MPLS Service Delivery Platforms" (see Ethernet/MPLS Growth Up for Grabs). Adding features such as ACEIS to the 7750 isn't impossible but is time consuming and risky, he writes.

Alwan responds by noting that ACEIS is shipping on the 7670 Routing Switch Platform and will, indeed, be added to the 7750. "We're taking that same set of guys that have done this on two platforms, and [creating] what I would say is a third-generation implementation on the TiMetra platform."

Alwan adds that he doesn't consider high availability to be an exclusive attribute of core routers; rather, he thinks it's something that edge routers will need to adapt as well.

— Craig Matsumoto, Senior Editor, Light Reading

For more on this topic, check out:

For further education, visit the archives of related Light Reading Webinars:

Page 1 / 3   >   >>
route_66 12/5/2012 | 1:29:05 AM
re: Alcatel Sticks to Core Plans "Alwan adds that he doesn't consider high availability to be an exclusive attribute of core routers; rather, he thinks it's something that edge routers will need to adapt as well."

This is something that most vendors don't seem to get. I'd make the stronger statement that HA is *much* more important on edge routers than in the core.

In the core, I have redundant equipment and circuits, and can rely on fast-reroute and my IGP to route around failures. On the edge though, the customer's circuit has to terminate on something, and that box needs to be HA, or the customer will be *down* throughout a hardware failure.

I've been disappointed over the years at all the companies that have failed to deliver working HA (N+1 eletronics, truly redundant routing and switching procs, etc.) for the edge (see, e.g., Unisphere ERX, Redback Smartedge) despite their promise.

That being said, I would agree with the skeptics that think building HA equipment into existing equipment like the 7750 is likely to fail. We've yet to see a vendor that can combine the hardware expertise to build a good HA platform with the software expertise to build a good router and MPLS services platform.

reoptic 12/5/2012 | 1:29:03 AM
re: Alcatel Sticks to Core Plans The Timetra product by all accounts is a long way from being finished and stable just for edge or multiservice features and now all the core elements and high availability pieces are being thrust upon it as well. Seems like something that will take years or will collapse under the weight of too many requirements. Hard enough to get some market share on the edge or the core with seperate focused teams let alone a giant effort. We have no strategy and no product that works yet so let's throw in the kitchen sink of requirements.
Stbl 12/5/2012 | 1:29:01 AM
re: Alcatel Sticks to Core Plans
The point made about HA earlier is an interesting one - in reality core routing from a feature requirement is fast becoming (or has become) a subset of edge routing. Some differences remain like performance and BGP scalability. Other protocols like LDP, HA, large route tables, etc. are increasingly expected edge requirements as well.

Bottom line, if you are building an edge router and it happens to have strong performance and a good BGP implementation you can likely play in many "core" applications. On the other hand, a product designed for the core will find it more and more difficult to meet edge requirements, particularly as IP/MPLS networks evolve to multi-service.
coreghost 12/5/2012 | 1:29:00 AM
re: Alcatel Sticks to Core Plans Whats happening at Alcatel is that Mr. Timetra
doesn't want any rivals or rival products within
the company. Its either going to be built by
"his" people or its not going to get built.

For the moment, he has the confidence of Paris
and until that confidence starts to slip, alcatel
will not seriously consider building a core router.

Its a very stupid and short-sighted strategy
thats more about someone building and protecting
a power-base than it is about business or

Timetra is not designed to be a core router. Its
good for what it is and it will fail if they
try and push it into a role it isn't meant for.
firstmile 12/5/2012 | 1:28:59 AM
re: Alcatel Sticks to Core Plans core,
Ridiculous statements. If Basil and the 7670 (newbridge guys) and Timetra guys cannot build this product for Alacatel, then THERE IS NO ONE ELSE IN ALCATEL THAT CAN EITHER. Do you know Alcatel at all? Do you think that there is some secret group of Cisco refugees hiding in Belgium? Maybe you are hiding in Belgium? :-)
And yes, he currently has the confidence of Paris, but he will fall out of favor. Unlike the rest of the gang in Paris, he is not French, and he is a very wealthy entrepreneur, AND he lives in California. With those three strikes against you, it is tough to have a significant senior executive carrier path in Paris.
Tony Li 12/5/2012 | 1:28:56 AM
re: Alcatel Sticks to Core Plans
I agree completely that the edge feature set is much richer than what is necessary in the core, but without performance AND scalability, an edge product isn't going to play in the core. However, a core box that is flexible and can add features over time has a hope of migrating to the edge. This seems lost on some folks.

Net: I don't think that there's a clear ordering between core and edge. It's very much in danger of becoming a lattice, where no migration is possible. That would be unfortunate for the industry.

digerato 12/5/2012 | 1:28:55 AM
re: Alcatel Sticks to Core Plans <<if (newbridge="" 7670="" alacatel,="" alcatel="" and="" basil="" build="" can="" cannot="" either="" else="" for="" guys="" guys)="" in="" is="" no="" one="" product="" that="" the="" then="" there="" this="" timetra="">>

ROFL -- very well put.

<<he a="" and="" california="" entrepreneur,="" french,="" he="" in="" is="" lives="" not="" very="" wealthy="">>

Not only that, he didn't go to one of the Grand Ecoles so doesn't have the right "old school tie". And rumor has it that he doesn't like Johnny Hallyday either :-)

But... wait a minute. Tony's post reminded me that Roland Acra from Procket is now out of a job and
* He's French
* He did go to one of the Grand Ecoles
* Don't know about they Hallyday question

What a great rumor possibility -- Acra goes to Alcatel and replaces Basil! Only two problems:

* Roland lives in CA
* He's very wealthy


6381891 12/5/2012 | 1:28:54 AM
re: Alcatel Sticks to Core Plans rumors are that alcatel is not selected in a core RFI for FT.. so on their own home ground....that will raise some questions within alcatel Paris
fredfrenzy 12/5/2012 | 1:28:54 AM
re: Alcatel Sticks to Core Plans There are of course certain golden handcuffs to hold basil et al for a bit yet.....they have got to focus on delivering the edge device they promised. No distractions like trying to break into the supercore duopoly. They're doing the right thing - like any newbie, focus, focus, focus.

That of couse doesn't mean that one more euro-cratic impatient purge won't happen.
beltway_light 12/5/2012 | 1:28:52 AM
re: Alcatel Sticks to Core Plans
> but without performance AND scalability, an edge
> product isn't going to play in the core.However,
> a core box that is flexible and can add features
> over time has a hope of migrating to the edge.

Isn't performance/scalability sort of exclusive
to the 'flexible' design ? I think either side
has problem to really cross the line to the other.
Page 1 / 3   >   >>
Sign In