x

Big Dig KOs Light Reading

Allegiance Telecom Inc. (Nasdaq: ALGX) is blaming construction workers in the Boston area for crushing an entire fiber duct belonging to the service provider Tuesday, causing massive disruption to hundreds of businesses up and down the eastern seaboard.

Allegiance, a national local exchange carrier (NLEC) serving 36 states across the U.S., initially blamed Level 3 Communications Inc. (Nasdaq: LVLT) for the outage, which disabled Light Reading -- and its sister publications Byte and Switch, Unstrung, and Boardwatch -- for most of the business day. The Boston Globe Website was also down all day, as was Infonetics Research Inc. and many other firms in the Boston area.

After a day full of finger-pointing, Allegiance finally admitted that the problem was theirs to solve and explain. "It turns out it wasn’t Level 3, or MCI, or Genuity Inc., as we had initially thought, but a giant piling that fell on our mega POP at 450 D Street in Boston,” says Jerry Ostergaard, director of public relations at Allegiance. The piling apparently cut four OC48 links.

Ostergaard was unable to confirm who was responsible for the damage. “They are saying it’s an outside contractor working on the Big Dig, but we aren’t sure.” He was also unable to say when the fiber cut is likely to be repaired. “It could be another couple of hours,” he said at 2:30 P.M. EST. [Ed. note: Light Reading came back online at 5:15 P.M. EST.]

Boston is in the process of building a vast underground roadway system, known as the Big Dig, which is costing billions of dollars (the latest estimate is $14.6 billion for the 7.5-mile highway) and has been responsible all manner of chaos in the Boston area, for which some have nicknamed it "The Road to Hell."

The outage will no doubt cost Allegiance big bucks, and the service provider is already in financial hot water. "A big swath of people have been affected by this," says Kevin Mitchell, an analyst with Infonetics Research. "It’s not good timing for the company."

Allegiance was supposed to reduce its loans to an agreed $645 million by April 30, but it failed to do so. Miraculously, it won a reprieve from its lenders, which extended the deadline until May 15.

"They are in a tight spot -- after paying $5 million to extend negotiations with the bank for this extra time, it still doesn’t look good,” says Cannon Carr, an analyst with CIBC World Markets. “It’s very difficult to do, but it’s in the banks’ best interests to make this work, as it’s not clear how much the company is worth at this point.” He adds that a likely possible outcome is a bankruptcy filing.

Allegiance’s share price rose 14 percent today, reaching $0.35.

— Jo Maitland, Senior Editor, Boardwatch

Page 1 / 3   >   >>
alchemy 12/5/2012 | 12:04:03 AM
re: Big Dig KOs Light Reading Yup- but how far can you sling a GE? Are there any native GE services running across a regional ring right now?

The MSOs routinely point gigabit ethernet from their Cisco routers into dark fiber. It's regional but in no way, shape, or form is it a ring. That's why 802.17/RPR over gigabit ethernet is so interesting to them as a MAN technology. They expect to get the 50 mSec protection switching of Sonet without shelling out the large bucks. So what if it degrades 50% on a fault? It may be held together with duct tape and paper clips but that's the current thinking on the future access network for the masses.

Since most GE are basically hop-to-hop, it would be too expensive to regen at each node in the link- sooo, would it not make sense to have a low cost, photonic routing layer?

As a TDM guy, that's what I said about ATM for years. Silicon improvements rendered my opinion obsolete.
brettnem 12/5/2012 | 12:04:05 AM
re: Big Dig KOs Light Reading Actually, I'm refering to a "transparent photonic switch" or an "OOO" device. That what I typically think of as OXC. I don't think the OXC replaces the huge bays of ADMs.. perhaps something like the CoreDirector does (ok, that could be an OXC of sorts..).. but I think a good technology that has replaced the huge bays of ADMs is DWDM and 3R transponders. Which I think is a better solution than the OEO OXC or ADMs.. But of course, that depends on your solution.

The basic point is OOO switch ports are rediculously expensive (both in $$$ and in dB) and simply don't do much more than a fiber panel. Especially when you lack an integrated control plane!

a Hybrid would be a neat device. I used to call the O(O/E)O interfaces "semi-opaque". I'm not sure if anything like this exists anymore.. Something like what Cinta had (very cool dead box).

Once again tho.. if you don't have an integrated control plane (all the way to the access device) then it's not really a lot more useful than a patch panel.

Remember, if an OOO photonic switch is going to cost you 10+ dB, you are going to have to pay for that port, plus all those amps, etc, etc.. I just don't see what you get for all that "cost". But hey, I realize that I personally may just not have come across the right application for it yet.

BTW, I am definately biased here.. In my last job, I spent my first 2 months there convincing the exisitng engineering team that a network of OOO switches was unnecessary ("useless" I think was the actuall explaination for our application..)
-Brett


trixie 12/5/2012 | 12:04:06 AM
re: Big Dig KOs Light Reading OXCs are really only good for very fast bandwidth provisioning (seconds to minutes) and for 1:N protection. Everything else just makes it look like an overpriced fiber panel (which also requires power, AC, OSS integration, and a SKILLED TECHNICIAN) Typically fancy rerouting of 1:1 or 1+1 circuits are better handeled by SONET or the various data routing protocols. Also, not to mention, there are limited [successful] applications for sub-hour provisioning. I think the last few years has been dedicated to reducing provisioning times from months to days. Which is a different application than.. say for example.. ACME bank does a major backup every friday from 3am to 4am and needs it's backbone beefed up during that time from OC-3 to OC-12. That requires a lot of hardware functionality that simply isn't very well implemented yet (not cross-platform at least..)

I agree- OXC's are money hogs re OEO interfaces, BUT, they are still leaps and bounds ahead of the legacy solution, which was counltess bays of ADMs, DCS and terminals.

I was talking more about a photonic switch than an OXC, although a hybrid would probably be the best overall solution.

I would agree that controkl pmane issues will be a barier for some time to make this a reality, but there are ways around it now, although more tedious to implement.
trixie 12/5/2012 | 12:04:06 AM
re: Big Dig KOs Light Reading Sorry for the newbie html experiment... italics [i]italics[/i]

Yeah, if it's SONET. It's cheap if it's GIG-E.


Yup- but how far can you sling a GE? Are there any native GE services running across a regional ring right now?

I always assumed they were packed up in a SONET frame, and sent merrily along the way, if they had some distance to travel.

Seems to me that GE services packed up in a "skinny SONET"- no ring protection, etc, just overheads for DCC- would lend well to an interesting wavelayer architecture in the network. Or just a GE wavelength itself on a wavelength. Since most GE are basically hop-to-hop, it would be too expensive to regen at each node in the link- sooo, would it not make sense to have a low cost, photonic routing layer?
brettnem 12/5/2012 | 12:04:07 AM
re: Big Dig KOs Light Reading Router ports ARE expensive.. But a OXC network + the costs of OSS integration, installation, truck rolls.. it's a lifetime commitment.

If I have to chose to add $100,000 in router ports, using my same hardware and chassis VS. adding a new major component to my network, sending people out to install it, integrating it into my backoffice OSS, teaching people how to use it.. I'd definately go for the router port. The OXCs are just too expensive and complicated and from MY point of view, simply don't add enough. Heck, if they did.. well.. you know the story...

Now duping the signal at the optical layer, like a simple headend bridge, tailend switch seems like a really good solution to this problem in my head. But then again, how big is your network.. Is it point to point? ring? mesh? Doing that kind of protection leaves your with some lack of control and capability to reconfigure the network. That, however, might be completely acceptable.. btw, I have used this method myself several times!

-Brett
dljvjbsl 12/5/2012 | 12:04:07 AM
re: Big Dig KOs Light Reading
For most, the preservation of active audio streams during a major fault may not justify the economics and requiring customer actions to reestablish those streams is probably acceptable.


It is a standard requirement in current systems that active calls will persist across faults. Recall that voice is used for life-critical applications.
brettnem 12/5/2012 | 12:04:07 AM
re: Big Dig KOs Light Reading Ok, here's my beef. I'd consider myself a next-gen engineer. I do all that fancy Cisco 15454 SONET stuff and some slick DWDM stuff in the metro. However, when it comes to voice and data I'm pretty traditional. I haven't seen a real compelling plan to combine voice and data. The required SLAs are just very different. Sure there are plenty of technologies that can do a good job of combining them. But in the end all, do you really save any($)thing? If you are the network provider (ie, own the fiber and components) bandwidth isn't that big a deal (a slight exageration). If you need to use two seperate circuits, one for data and one for voice to guarentee excellent quality, go for it.

I also can see the benifit a previous poster mentioned regarding a distributed protection mechinism. However, it 1+1/1:1 protected systems with OXC, the backup path is pre-decided (to ensure < 50ms switch) and there is really no distributed protection. In fact, there is distributed exposure now. It depends on the implementation of course. It certainly can be done, but has anyone made a dynamic reconfigurable OXC mesh that can FIND and provision a new end-to-end route in < 50ms?

So for data networks, I'd say OSPF, or something along those lines.. For Voice.. well, SONET does a pretty darn good job..

Look, when it comes to PROTECTION, you have REDUNDANCY and GEOGRAPHIC DIVERSITY. Sometimes people say a route is DIVERSE and that simply means you have redundant fibers, but they run in the same DUCT. It is important to understand how your provider defines these things. I like to call it "GEOGRAPHIC DIVERSITY" so there is no question that the redundant routes follow different paths.. sometimes even go so far as to say that at no time do the paths cross or come within 30' of each other. Or at LEAST one large persons arm span apart (ie, left hand can't disconnect work while right hand is disconnecting protect). Of course, that MIGHT break down a bit towards the terminating rack..


-Brett
trixie 12/5/2012 | 12:04:08 AM
re: Big Dig KOs Light Reading What I'm saying is that without progressive thought, backed by real actions, and based upon principles, a society's future becomes little more than a theocracy led by fundamentalist zealots who have no respect for human rights. That's a recipe for disaster.

We're better than that.


Hmmmmmmm....

I was just looking for ways to discuss cost-effective highly available services. Perhaps we could start a seperate metaphysical thread to discuss these issues...
rjmcmahon 12/5/2012 | 12:04:08 AM
re: Big Dig KOs Light Reading I was just looking for ways to discuss cost-effective highly available services. Perhaps we could start a seperate metaphysical thread to discuss these issues...

Fair enough. Though the issues are more intertwined than most realize. Sorry for the rants.

One suggestion is that we begin solving the high availability problem by reducing the monthly IO costs and increasing bandwidth and connectivity by orders of magnitiude. Establishing a competitive market, for both equipment vendors and service providers, such that modern technology will be deployed as our basic infrastructure, would also help.
rjmcmahon 12/5/2012 | 12:04:09 AM
re: Big Dig KOs Light Reading truly a service could always be available, and still be complete garbage

One man's trash is another man's treasure. The number one US export, if measured by weight, is recycled newspaper of which we all throw away daily. It goes to Asia where they use it in the production of all kinds of things.

The point is, it's not for the middleman to judge the worthiness of any service. Let the transactions occur without the interference.

Are you saying the CLEC/NLECS are just hemmoraging the industry? An outright monopoly would be the better play?

Natural monopolies exist. Regulatory capture exists. The CLECs have failed. The 1996 Telco Act has been little more than disgrace to the investment community. The FCC fails us daily.

What I'm saying is that without progressive thought, backed by real actions, and based upon principles, a society's future becomes little more than a theocracy led by fundamentalist zealots who have no respect for human rights. That's a recipe for disaster.

We're better than that.
Page 1 / 3   >   >>
HOME
Sign In
SEARCH
CLOSE
MORE
CLOSE