Optical/IP Networks

Report Heralds IP Doomsday

More sour news on the telecom farm -- or just another Chicken Little?

A widely circulating joint research report from McKinsey & Company and J.P. Morgan & Co. (Nasdaq: JPM) says a decline in the growth of Internet Protocol (IP) network traffic is partially to blame for the ongoing shakeout among telecom service providers.

The report predicts that IP traffic growth will slow from its recent 200-to-300 percent growth rate to a 60 percent annual increase by 2005, mainly due to a slowdown in the growth of Web-page traffic. This slowdown, the report predicts, will hurt backbone carriers the most, as better management at the network's edge and aggressive pricing eat into pure IP transport revenues.

While the report still sees overall IP traffic revenues growing from $13 billion in 2000 to more than $56 billion by 2005, it also predicts that network-edge traffic aggregation (through Web hosting and content delivery networks) could steal away as much as $6 billion per year in transport revenues from backbone providers by 2005.

"Industry players that were formed around what was believed to be an insatiable drive toward Web-centric traffic and revenue growth," the report claims, "will soon have to begin rewriting business plans, if they haven't already, to pivot toward enterprise customers -- a goal for which many will find themselves ill equipped."

"These findings are good news for backbone carriers with rich customer relationships," says Paul Roche, a principal at McKinsey who helped prepare the report. "But for single-use or Web-centric providers, this is trouble."

Roche says McKinsey and J.P. Morgan conducted a wide range of interviews with service providers to get the numbers for their study, which contains a number of caveats attached to its modeling and predictions.

Though some observers question some of the report's specific numerical findings, they tend to agree with its conclusion that telecom providers need to develop specialized IP services to survive the current telecom nuclear winter. For example, the report predicts that services such as virtual private networks (VPNs) are the key to survival. Similar thinking was no doubt behind Juniper Networks Inc.'s (Nasdaq: JNPR) decision this week to add more IP functionality to its routers (see Juniper Mounts Pulpit for IP Services).

"People who are only selling pipes are the ones who will die," says Faizel Lakhani, vice president of network solutions at Caspian Networks, which is building network core routers.

The report predicts that enterprise-centric traffic will account for the lion's share of IP-based revenues going forward, a trend that is already being addressed by some equipment and service providers that are focusing on providing managed services instead of pure bandwidth. The report also says that backbone providers will face "continued economic dislocation" while network edge providers will grow in importance as more IP traffic gets managed and controlled closer to the actual customer.

Several industry observers questioned the basic premises of the report, saying both that slowing growth was somewhat inevitable (since traffic levels are already high) and that some IP traffic patterns, particularly outside the U.S., are still in the formative stages.

Vint Cerf, one of the Internet's early pioneers and now a senior vice president for WorldCom Inc. (Nasdaq: WCOM), questions both the report's current and future IP traffic estimates. "I doubt traffic ever grew at 300 to 400 percent per year," says Cerf, responding via email. "I think it has been more like 80-100 percent [annual growth] which is likely to continue, especially as broadband is more readily available and more widely deployed."

Others questioned some of the data.

"It's not clear if the report takes into account Internet traffic that will come from places like India and China, which are wild cards when it comes to predictions," says Ron Westfall, a principal analyst with research house Current Analysis.

Many industry players, however, are already embracing the changes discussed in the report. Sprint Corp. (NYSE: FON), for instance, formed a separate business unit last November that is aimed at providing more managed services, on top of hosting and colocation, to enterprise customers.

Level 3 Communications Inc. (Nasdaq: LVLT), which has an established colocation business, recently announced plans to support voice-over-IP services; and earlier this week, Cable and Wireless (NYSE: CWP) snapped up content-delivery provider Digital Island (Nasdaq: ISLD) for $340 million, helping confirm the report's prediction that consolidation will be rampant as backbone players and ILECs seek to add functionality -- while the single-service players collapse.

"We expect there to be a lot more M&A as carriers try to add scope," Roche says. "And we're pretty sure that the absolute number of backbone providers is going to shrink, going forward."

Though Roche says that the backbone consolidation may also hurt network core equipment providers, Caspian's Lakhani disagrees with some of the report's traffic-decline figures, saying that his company's sources see both edge and backbone IP traffic continuing to increase.

"It seems like there's some confusion [in the report] over what's edge growth versus core traffic growth," Lakhani says, noting that because the path of IP traffic is largely unpredictable, the core capacity must increase at a multiple of the growth at the edge to handle unforeseen events like broken links or spikes in traffic.

Lakhani, who says that Caspian obtains its own network-traffic figures from its non-disclosure agreements with "19 leading ISPs," says that corporate IP traffic may be increasing even faster than the report predicts.

"Corporations are moving from T1s to T3s far more quickly than anyone has anticipated," he says. Another increase will come as pure IP services start to offer the quality-of-service attributes of current WAN services.

Though not yet available on the Web, the full 112-page report (titled "IP! -- How Changes in the Internet Are Disrupting the Telecom Services Industry") can be obtained from either McKinsey or J.P. Morgan, Roche says.

- Paul Kapustka, Editor at Large, Light Reading

Page 1 / 4   >   >>
flanker 12/4/2012 | 8:25:17 PM
re: Report Heralds IP Doomsday I've dealt with McKinsey before. These guys are not information age gurus, ESPECIALLY on the carrier network side.

Talk to PWC, RHK, KPMG, IDC, Forrester, Telegeography...

No one is talking about declining traffic.

Secondly, there are network sizing issues relating to capacity utilization. Historically, carriers have bought more capacity than they need to handle surges in traffic. Overhead on carrier networks seems to always end up around 3x average load, even though it is not a targeted figure.

Again, if you want to offer five-9 reliability, you have to overbuild and over-provision.

These sort of factors do not lend themselves to the conclusion that data revenues will decline. On the contrary, data is a core ILEC/IXC offering given the collapse in ILD/LD voice revenues.

Finally, regardless of whether the traffic is metro or long haul, you still need to provision for capacity spikes ON BOTH SIDES to avoid a bottleneck/network outage.

I respect McKinsey, who probably did a ground-up study of traffic flows, but they are not operators and don't think like operators. You have to factor five-9 reliability and it means provisioning circuit capacity equal to a substantial multiple of normal traffic flows in other to assure throughput and avoid network crashes.

switchrus 12/4/2012 | 8:25:14 PM
re: Report Heralds IP Doomsday Old Business Model:

Hi IGÇÖm a big TELCO, I have really cool commercials about Pins and Dimes. You want to move large volumes of data from point A to point B ? When, how much, on Tuesdays ? ? Have I got a deal for you, please bend over.

New Business Model

IGÇÖm a big company that produces lots of data every day. Once a day I want to back the truck up, load my data on it and move it to my warehouse in Kansas. IGÇÖve got a pretty good engineering team and some really tight fisted green eyeshade folks. I can roll my own network, I can buy time on a network I can buy a lot of little trucks itGÇÖs all about what the green eyeshade people tells me works. I donGÇÖt need Pins or Dimes and all that overhead.

You know, Dinosaurs did eventually die off.
jmd 12/4/2012 | 8:25:14 PM
re: Report Heralds IP Doomsday Is there really some new insight here? This article is indicating that this report is about jumping on the bandwagon of whatGÇÖs already been said.

Bandwidth is commodity and is no longer a basis for revenues that scale. Webpages are not the future of the Internet. IP traffic management at the edge (efficiencies above layer 4) can be leveraged. Service providers want something that they can generate revenue with. Enterprise is back. The consumer market is not cracked. M&A in the backbone is needed.
flashlight 12/4/2012 | 8:25:13 PM
re: Report Heralds IP Doomsday Ask those Mckinsey guys about the report that they did on the Celluler industry for AT&T?
cfaller 12/4/2012 | 8:25:09 PM
re: Report Heralds IP Doomsday flanker said:

"I've dealt with McKinsey before. These guys are not information age gurus, ESPECIALLY on the carrier network side. Talk to PWC, RHK, KPMG, IDC, Forrester, Telegeography...

No one is talking about declining traffic."

So what if PWC, RHK, etc. aren't talking about this? They didn't predict this telecom crash that we're in the middle of, did they?

McKinsey based this report on conversations with service providers, not with PWC, RHK, etc. For that I would give the report more weight. The idea that other consultants know more about the service provider sector than service providers themselves is ridiculous, laughable, flat wrong, etc.

I'm not ready to completely embrace this report, but I'm not going to dismiss it out of hand either, and I'm certainly not going to base that dismissal on other consultants' reports!
fk 12/4/2012 | 8:25:09 PM
re: Report Heralds IP Doomsday If we are going to attack the report, let's attack it for what it says. It didn't say anything about traffic declining. It said the rate of growth of traffic would decline. That's a significant difference. If someone told me the amount of traffic would decline, I'd think them a kook. If someone told me that traffic wasn't going to continue growing as fast as it has the past few years, I'd listen to what they had to say. I think it's a matter of the math of big numbers. Traffic can't grow 300-500% per year every year very long before the math simply gets out of hand. I think slowing to sixty percent yearly growth is a perfectly reasonable conclusion within 5 years. Does anybody really think that traffic will grow to infinity?
LightCycle 12/4/2012 | 8:25:07 PM
re: Report Heralds IP Doomsday > how do they compare against Juniper's products?

Unisphere is (for now) a one box shop - the ERX - and its not so much a router as it is an IP services switch, or what might be termed a DSL aggregation platform (ala Redback).

Ie. it does not compete in Juniper's primary space.

In any case, the underdogs will almost universally show market share gain (assuming they have something worth selling). One box today and two boxes tomorrow = 100% growth.

Belzebutt 12/4/2012 | 8:25:07 PM
re: Report Heralds IP Doomsday In this article it states that Juniper and Unisphere both gained market share despite the shrinkage of the router market:


I'm not too familiar with Unisphere's products, aside from the marketing stuff (http://www.unispherenetworks.c..., how do they compare against Juniper's products? Seems like they only support up to OC-12...
wimchatta 12/4/2012 | 8:25:06 PM
re: Report Heralds IP Doomsday Some observations on the basis of the LightReading article, and having worked at McKinsey (again with Morgan) on a previous study of Backbone infrastructure before reentering the startup world:

i) McKinsey tends to do very thorough analysis with the data available. Morgan provides a quantitative, "what does this mean to the financial community" perspective.

ii) Service Providers are clearly the closest to the traditional customers, and hence a good source of information on what is expected in terms of developments among those customers. They are unlikely to have bankable opinions of radically new market developments, as Christensen would indicate in his treatise on disruptive technology.

iii) Between McKinsey and Morgan, all the significant service providers are likely to have been contacted, including the ones with newer optical infrastructure. This report ought to be as reliable (if not more) as any published by other firms. RHK is possibly the only source whom I would give more weight to, as they also listen closely to service providers and have been doing so for many many years.

That said, the report has come after a period during which few of the "sources" could have been confident about their view of the market. Hence the timing of publication is such that the information available for the report is questionable. The conclusion would have been very different if the report had been done a year ago, and is likely to be different if it is done a year from now. But that will be likely be the case for any other industry report published by anyone else.

skeptic 12/4/2012 | 8:25:06 PM
re: Report Heralds IP Doomsday I agree with parts of the report. Traffic cannot grow as fast as it was but 60% growth is still not something to dismiss as a financial disaster.

What the report didn't say is that the operational and equipment pratices at many of the large carriers are a large part of why their profits are so bad.

What this really means is that the marginal players are going to be out of business soon. The worst designed (or undesigned) carrier networks are going to go out of business and that the two places to make money (IMO) are:

1. reducing the cost of the network as a whole
2. offering value-add services at the edge that people will pay for.

With 60% growth a year, that still means replacement of entire networks on a really frequent basis. And while the carriers can hold off spending for a few months, they have to spend to keep up with that growth or exit the game.

Page 1 / 4   >   >>
Sign In