x
Optical/IP

Force10 Strikes at Sonet

Force10 Networks Inc. is muscling into Sonet (Synchronous Optical NETwork) and SDH (Synchronous Digital Hierarchy) territory, and is aiming to wrest service provider accounts from Cisco Systems Inc. (Nasdaq: CSCO) and Juniper Networks Inc. (NYSE: JNPR). (See Force10 Adds.)

The company is releasing a 10-Gbit/s card -- four ports of OC48 -- for its E-series switches that's being targeted at older telecom networks.

OC48: Isn't that yesterday's yawner news?

The kicker is that Force10 claims it can give carriers that 10-Gbit/s card for half the cost of an equivalent blade from Cisco or Juniper.

The key is that Force10's E-series is an Ethernet box. It's been designed to provide a cheaper alternative to a full-blown IP router, and its interfaces are less expensive than the usual Sonet/SDH fare.

In going to Sonet/SDH, Force10 is using its usual Ethernet card, adorned with an interface that wraps Sonet/SDH traffic into an Ethernet envelope. Internally, the E-series then switches the traffic as if it were Ethernet, and translates it back to Sonet/SDH on the way out. Since the 4xOC48 card is Ethernet-based at its heart, it sells for $25,000 -- half the price of a comparable Juniper T-series card, Force10 claims.

In case you're wondering, Force10 didn't do a straight OC192 card because the multiple-OC48 market is larger. Plus, the 10-Gbit/s Ethernet WAN PHY interface is already defined to handle Sonet/SDH -- so, technically, Force10 already supports OC192.

This kind of thing never worked before for two reasons. First, Ethernet wasn't trusted enough in carrier networks. Jitter and latency weren't up to the standard set by Sonet/SDH. Ethernet has been getting beefed up, though, to counter this point.

Second, and more importantly, carriers weren't that concerned with the price of equipment. Sonet/SDH gear was expensive and worth it, because carriers prized five-nines reliability over cost.

Force10's hope is that the latter factor is melting away. Carriers need to make new services profitable while making the most of the installed Sonet/SDH infrastructure. Force10 is hoping they'll make do with less than a full-blown router if it means saving some capital expenditure.

"The price they can charge their customers is declining, and the price of that bandwidth isn't -- classic squeeze," says Andrew Feldman, Force10 VP of marketing. "For the first time, carriers are willing to trade features for price."

And there are spots in the network where that line of thinking works, according to Zeus Kerravala, an analyst with Yankee Group Research Inc. "If what you're looking for is something that's big, fast, and passes dumb packets," he says, "from an economic standpoint, it makes a lot of sense."

That doesn't mean it works everywhere, though. Cisco and Juniper are well ensconced in the world's largest carriers, and Force10 isn't likely to change that.

"To say this is going to change things overnight is a bit radical," says Kerravala. "It's not the death of routing yet."

The big guys agree. Well, sort of; Juniper declined comment. But the Juniper line of thinking probably echoes a Cisco spokesman's comment, namely, that carriers "are willing to pay for the feature set we provide."

But the Yankee analyst reckons there might be some slow creep towards Force10's way of thinking: "You'll see a gradual migration to this. What it's going to take is a couple of big-name carriers to bite."

But they probably won't be the conservative U.S. RBOCs -- they'd be the last of the carriers to try Force10's brand of Sonet, reckons Kerravala.

"Force10 and Foundry are pretty small -- you add them together, and you get maybe $500 million," Kerravala says. "But that's how change happens. One customer at a time, and one vendor at a time. That's how Juniper started."

Force10's Sonet/SDH card is in beta testing now, with general availability slated for August.

Meanwhile, Force10 says it's still enjoying revenue growth of 100 percent per year, and it's outgrown its original headquarters. Last week, the company finished a move to North San Jose, Calif., occupying a grandiose building next door to Redback Networks Inc. .

Which brings to mind... Whatever happened to that IPO? (See Sources: Force10's Prepping Its IPO.)

Officials won't comment, other than to say it's still in their plans. In general, sources note that between Sarbanes-Oxley rules and the ongoing stock-options scandal, the IPO process has gotten more difficult.

— Craig Matsumoto, Senior Editor, Light Reading

Page 1 / 2   >   >>
Pete Baldwin 12/5/2012 | 3:47:28 AM
re: Force10 Strikes at Sonet So, how big a deal is this, really? In, say, five years, could Cisco and Juniper see some market-share losses (service provider market) from Ethernet-based boxes? And if so, what should they do about it?
Pete Baldwin 12/5/2012 | 3:47:27 AM
re: Force10 Strikes at Sonet Re: I was expecting this title to be in reference to open source routing code that runs on common equipment such as that just announced by Vyatta.

Ah, no .. sorry. That's a separate issue (and a separate story that will be forthcoming).
deanfro 12/5/2012 | 3:47:27 AM
re: Force10 Strikes at Sonet A new line card by a startup kills Cisco! Ok, I will let you get away with some dramatic overstatement as you are trying to spark debate. Perhaps a better title was death to overcharging for line cards.
The question needs to be what am I actually getting for 2 times or greater the cost with a traditional router interface beyond a Ethernet switch heritage box. Traffic management likely, and carrier grade?? maybe, what else. SPs need to think about whether they are getting value for the premium they are paying.

To your question of timing there is also the point of how important PoS is in 5 years.
lightreceding 12/5/2012 | 3:47:27 AM
re: Force10 Strikes at Sonet I was expecting this title to be in reference to open source routing code that runs on common equipment such as that just announced by Vyatta.

x25 12/5/2012 | 3:47:26 AM
re: Force10 Strikes at Sonet I would treat cisco & juniper very differently. Cisco has a broad portfolio & has handled previous market transitions well (OK..layoffs & some in fighting but successful). Thus I think they'll manage with the growth of L2 backbones just fine.

Conversely juniper has a more concentrated product line & less of a history managing large customers. This doesn't mean they'll lose out to cisco (i.e. they can move faster if they desire) - its just they have no track record.
lightreceding 12/5/2012 | 3:47:26 AM
re: Force10 Strikes at Sonet well the world of networking is moving to Ethernet everywhere. Ethernet owns the LAN and the data center and is moving to the WAN. So the comments that the F-10 guys make may be valid, but that doesn't mean that F-10 will put a dent in Cisco so fast. I see this interface as in interim play to help transition networks. That transition will continue to happen but there is time for Cisco to catch up and other players to enter the market.
tmc1 12/5/2012 | 3:47:26 AM
re: Force10 Strikes at Sonet "Force10 and Foundry are pretty small -- you add them together, and you get maybe $500 million," Kerravala says. "But that's how change happens. One customer at a time, and one vendor at a time. That's how Juniper started."
--------------------------------

nice quote sparky, but also totally incorrect regarding juniper. juniper was funded and encouraged by the customers of cisco and many of cisco's competitors to offer a second serious routing vendor (after the fall of wellfleet). not at all like this case. foundry and f10 are just two of many, many ethernet switching companies already existing. nice hyperbole though!
jcupit9999 12/5/2012 | 3:47:25 AM
re: Force10 Strikes at Sonet One last point - this is an approach that has been tried before by other vendors at the DS-3 level - anyone remember Seranoa Networks?

Again - the operational issues precluded the use of this device in the carrier environment which could convert 12 DS-3s to a GE interface. The primary benefit of the product was that it gave carriers a weapon to drive 12-port DS-3 interface pricing downward.

jcupit9999 12/5/2012 | 3:47:25 AM
re: Force10 Strikes at Sonet While it is true that there may be economies of scale to be realized by utilizing the F10 approach, it appears that their approach would create an issue of protocol conversion at the card level, which would be deemed as an enhanced service from a regulated carrier perspective - this is a non-starter inside of the Tier 1s - and many of the Tier 2/3s are already beneffiting from the lower OC-48 interface costs provided by the micro MSPPs (White Rock, ADTRAN, etc.).

I am not aware of many Tier 3s who would require the capacity of the F10 box in support of its production environment - there may be 5 to 7 Tier 2s who could use the capacity - but one wonders if the product manager who created the product specification for this interface really understood the day-to-day operational and provisioning requirements of the market that this interface was created for.
sgan201 12/5/2012 | 3:47:24 AM
re: Force10 Strikes at Sonet Hi,

Do anyone realize how much bandwidth do you lose when you do GigE framing at the WAN?? Especially if you have a lot of small 64 bytes packets?? Or, are we assuming that bandwidth is free at OC-3/OC-12 level??

Dreamer
Page 1 / 2   >   >>
HOME
Sign In
SEARCH
CLOSE
MORE
CLOSE