Juniper Networks Inc. (Nasdaq: JNPR) is announcing what it claims will be the first available 100-Gbit/s Ethernet interface for a router, targeting the kinds of requirements Verizon Communications Inc. (NYSE: VZ) has been talking about.
In fact, Verizon features prominently Juniper's press release: The carrier has been pining for a more efficient -- and cheaper -- way to deal with the multiple 10-Gbit/s Ethernet lines that populate its network. (See LR Live: Verizon Plots 100 GigE RFP and Verizon's Wellbrock: 100G Is Needed.)
Verizon has "a strong desire to bring 100-Gbit/s technology into the network primarily to simplify the topologies," says Luc Ceuppens, senior director of marketing at Juniper. "They have a lot of 10 Gbit/s now and can eliminate a lot of those links."
Juniper's 100-Gbit/s interface is going on the T1600 core router and is housed in a standard CFP module. It's intended, not for metro or long-haul reaches, but for the box-to-box connections within a data center or central office.
The T1600 can run 100 Gbit/s per slot without oversubscription, and its switch fabric has the capacity to operate without oversubscription even if all eight slots run 100-Gbit/s interfaces, Ceuppens says.
Juniper might be faster to market, though, as it expects the T1600's 100-Gbit/s ports to go into customer trials later this year.
Verizon is targeting a 2010 deployment for 100-Gbit/s technology.
Companies including Ciena Corp. (Nasdaq: CIEN) and Nortel Networks Ltd. have begun toying with the idea of pure 100-Gbit/s feeds. But carriers are saying they really need a 100-Gbit/s wavelength as a carrying vehicle for multiple 10-Gbit/s feeds, Ceuppens says. "They need to eliminate some of the inefficiencies of link aggregation." (See Ciena Sending 100GE Live and Nortel Shows Single-Slice 100GE.)
The cost of the 100-Gbit/s link is going to be "just under" that of ten 10-Gbit/s links at first, Ceuppens says. How quickly that price drops will depend on component availability. Still, that's an improvement over 40-Gbit/s Ethernet, where the price remains higher than that of four 10-Gbit/s links, he says.
Interested in learning more on this topic? Then come to Ethernet Expo 2009, Light Reading’s ninth conference and exposition covering the hot topic of Carrier Ethernet network technologies and services in North America. To be staged in New York, November 3 & 4, admission is free for attendees meeting our prequalification criteria. For more information, or to register, click here.
No, my organization is not likely to buy 100GE blades any soon (or at least until they become financially affordable).
As far as the list of JNPR customers, the press release is very clear - it says shipping starts this fall and that's when we will likely know who got hands on the new technology.
My gut feeling is that first units will end up at Tier1 operators, and it's highly unlikely they will speak about their experience at LR forums anyway.
Frankly, I am not that interested in who ships production units first, although it seems quite possible that Juniper will indeed be the leader.
My oersonal interest is to see who can drive the cost of 100GE to undercut OC768. I am also expecting the existence of 100GE to send 10GE routed port prices south of $10k.
Even if you personally don't like Juniper (for whatever reason), their work stimulates the new market and creates competitive pressure.
Hey... dude... No one is saying that there is no market for 100Ge... nor is any one saying where early adopters may use 100GE (read the thread) All i'm saying is who is buying it FROM JUNIPER today? No vendor NOR customer has voiced their opinion here.
The original statement was "Juniper claims 100-Gig first" - Do you work for Juniper? Are you an ealry adopter and or planning on buying 100G from Juniper?
Apparently you were not around at the dawn of 10GE.
The first 100GE applications will likely be in local interconnects - eq exchange points currently linked with NxOC768. Fiber type does not matter much within the facillity.
Interoperability also is not a big deal - most exchanges are single- or dual-vendor at the most. What matters more is the port price compared to 10x10GE or 2xOC768.
I guess we will hear customer names when Juniper ships in autumn (and they have a track record for shipping products on time). I wont be surprised to see DTs and FTs of this world to be the sole customers.
Picture will become much more interesting later in 2010, when more companies will deliver 100GE products and price wars will start brewing, Coupled with better optics, 100GE interfaces will surely become a de-facto LAN interconnect standard.
NOT - 100G on the client side folding in to a 10x GE core mesh? I'll leave you alone on that one... lol.
Phrase that how you wish...are you thinking of buying or are you a vendor thats trying to market yourself?
I think you and I know that showing gear at events like Interop is a LONG way from functioning products and customers actually buying it. Getting to the real questions, and basis for my comment - which no one has chosen to answer yet... who's buying it? What fiber can support it?
Even after the standards are ratified, it takes a long way before equipment from different vendors can de-facto interoperate.
So kudos to Juniper for making the first packet processing unit with 100GE optics.
Now the real question - how cheaper it is compared to 10x10GE bundle?
If there is significant saving from buying 100GE interface over 10x10GE, it will sell even if it works only between Juniper devices (T-series is the de-facto standard core router nowadays).
I am also pretty sure that Juniper will factor some protection policy for early adopters. Judging from impact of 10GE, i can't tell enough on how important the 100GE is to the industry... hope Juniper can drive the cost down quickly.
... This was in response to "Juniper Claims 100GE first" but since you decided to market yourself here:
1.) Key statement here: "As the IEEE P802.3ba standard moves toward finalization in 2010" (MAYBE).... Who's buying pre-standard technology today? How will more than 50% of the existing fiber in the carrier world support this today? Who's making the optics to support this?
2.) What third party can verify this claim? "Xilinx demonstrated in hardware (with no cheating!) the first commercially available single FPGA 100GE solution with our partners Netlogic, Sarance, Avago and Ixia at OFC in late March 2009." The fact that you say "in hardware" means thats it's half baked.
Just to update everyone on a few things regarding the state of 100GE:
(1) 100GE test equipment has existed for nearly 1 year through Ixia: http://www.ixiacom.com/solutions/higher_speed_ethernet_testing/index.php
(2) Xilinx demonstrated in hardware (with no cheating!) the first commercially available single FPGA 100GE solution with our partners Netlogic, Sarance, Avago and Ixia at OFC in late March 2009: http://press.xilinx.com/phoenix.zhtml?c=212763&p=irol-newsArticle&ID=1269187
(3) You can learn more about hardware-proven, low risk solutions for 100GE from Xilinx on both our production Virtex-5 TXT and now-sampling Virtex-6 FPGAs on our website here:
http://www.xilinx.com/products/virtex5/txt.htm
http://www.xilinx.com/6
Lots of hard work from many people, but it was a hoot.
The blogs and comments are the opinions only of the writers and do not reflect the views of Light Reading. They are no substitute for your own research and should not be relied upon for trading or any other purpose.
To save this item to your list of favorite Light Reading content so you can find it later in your Profile page, click the "Save It" button next to the item.
If you found this interesting or useful, please use the links to the services below to share it with other readers. You will need a free account with each service to share an item via that service.
Digg
Del.icio.us
Reddit
