• United States

Who says the Bells aren’t spending?

Mar 10, 200311 mins

Capital spending by North American carriers this year – expected to be $30 billion – might be only half that of 2001, but the four regional Bell operating companies will invest nearly three-quarters of those dollars on potentially lucrative data services to replace eroding voice revenue.

The Bells’ plans call for building nationwide backbones and points of presence to enter a long-distance market opened to them as they met demands of government regulations. Services on tap range from IP VPNs and IP Centrex to basic long-distance.

“These buildouts are as much defensive as offensive,” says Thomas Nolle, president of consultancy CIMI. “By 2005 or 2006, voice will not be capable of funding the public network.”

Detailed reports

• Verizon• SBCBellSouth  Qwest• Next-gen voice services• Peeking at the RFPs

Verizon on the edge

Verizon, the biggest RBOC, plans to spend at least twice as much as any other – $13 billion this year (down roughly 25% from the $17.5 billion average spent by Verizon in 2000 and 2001).

Much of Verizon’s spending is going to core and edge IP routers, a new generation of ATM edge switches, and higher-capacity and data-friendly optical gear.

“[The investment is] really now just a question of how to meet customer requirements and manage a migration of some of the services” into the new infrastructure, says Mark Wegleitner, senior vice president and CTO at Verizon.

For optical services, the carrier is utilizing wavelength division multiplexing (WDM) technology as a vehicle for a variety of service offerings. WDM increases fiber capacity and provides faster service provisioning than SONET, carriers say.

Services Verizon plans to deliver over wavelengths include traditional SONET-based TDM, such as voice; Gigabit Ethernet data; and storage-area network offerings based on Enterprise Systems Connection, Fibre Connection and Fibre Channel, Wegleitner says.

WDM is the transport mechanism for those services under Verizon’s recently announced Enterprise Advance initiative, he says.

“We’re seeing an emerging market requirement from enterprise customers for wavelength capabilities,” Wegleitner says. “What we’re doing with Enterprise Advance is going upchain into managed services using wavelength as the transport.”

Corporations are attracted to wavelength-delivered services because they require less fiber and are provisioned more quickly than traditional SONET-based services, observers say.

For IP services, Verizon has RFPs out for core and edge routers along with multiservice switches. In edge routing, Verizon reportedly is spending up to $300 million on next-generation IP service routers (more on the RBOC RFPs).

The carrier plans to build out metropolitan IP networks and scale up the IP capabilities of its backbone to construct a multiservice infrastructure for end-to-end packet services on a national or perhaps even an international scale.

This end-to-end IP network will support Verizon’s current DSL aggregation and Internet access applications, as well as IP VPNs for enterprise customers, Wegleitner says. Other applications, such as voice-over-IP and triple play – voice, video and data – will follow.

Multi-protocol Label Switching (MPLS) enables all of this. It is a required component of virtually all products that the RBOCs evaluate for next-generation buildouts. MPLS lets Verizon tunnel legacy but revenue-rich Layer 2 access services such as frame relay and ATM, and a newer Transparent Ethernet LAN offering, through an IP backbone that provides quality of service (QoS), traffic engineering and reliable failover, he says.

“MPLS is to be treated as the target,” Wegleitner says. “That’ll give us the ability to provide all the things at Layer 3 or at Layer 2, and do it all on a common infrastructure.”

All this activity, however, has escaped at least some Verizon customers.

“I have not heard anything from them in regards to [new] services,” says Robert Primavera, assistant vice president at Eastern Bank in Lynn, Mass. “One thing that I’ve been watching very closely is what they’ll invest in state-of-the-art technology that we need to converge our voice and data around.”

Uniform maker Unifirst similarly has not been privy to any new service rollouts Verizon might have in the pipeline. But the Wilmington, Mass., company has its own RFP out on the street for a frame relay service to interconnect its sites across the country.

Unifirst wanted to evaluate Verizon for that but the carrier did not respond to the RFP, says Doug Hogue, project manager for telecommunications.

“They basically decided it’s a national network and not to bid on it,” Hogue says. “The reason we were given is that they didn’t feel they could be competitive.”

That could soon change. One of the key reasons RBOCs are engaging in these next-generation buildouts is the expectation that they’ll be able to offer national long-distance service under Section 271 of the Telecommunications Act of 1996. That proviso states that the local exchange carriers can enter the long-distance market provided they open their local loops to competitors.

And a key motivation for RBOCs to offer long-distance is the requirement that Hogue described: the need for frame relay service nationwide. Currently, interexchange carriers AT&T, Sprint and WorldCom dominate that multibillion-dollar market.

“[The RBOCs have] really been shut out of the most lucrative part of the frame relay business, and that’s the long-haul portion,” says Curtis Price, an analyst at Stratecast Partners. “Section 271 gives them the opportunity to go head to head with those companies on the long-haul portion. When you think about how large that frame relay market is, stealing market share from the three [interexchange carriers] is going to be the low-hanging fruit that they go after.”

SBC’s data plan

SBC offers frame relay and ATM services in Texas, Missouri, Oklahoma, Kansas and Arkansas. But in anticipation of 271 relief, SBC’s $5.5 billion budget this year – about half of what it was in 2001 – includes deployment of core routers for a national IP backbone to support all its data services, including frame, ATM and IP VPNs (more details).

Many analysts say IP VPN tunnels eventually could replace frame relay and ATM permanent virtual circuits. SBC is deploying a new generation of edge routers that might anchor a new IP-enabled frame relay service that supports frame relay or ATM access to a routed IP VPN, says Ralph Ballart, vice president of broadband infrastructure and services at SBC’s Technology Resources unit.

MPLS makes this possible. SBC’s edge routers support RFC 2547, an IETF specification for the construction of MPLS-based Layer 3 IP VPNs.

“We’re obviously looking at pure IP capabilities that would be end-to-end IP,” Ballart says. MPLS would enable traffic engineering, class of service and voice over IP (VoIP) on the backbone.

MPLS also plays a role in SBC’s optical trials, buildouts and impending service rollouts. SBC has a metropolitan transparent LAN service project under way that will use MPLS for scalability by alleviating the need for Ethernet switches to know the media-access-control addresses of all computer systems attached to the transparent LAN network, Ballart says.

So SBC will be invest in new Ethernet switches to overlay older SONET systems that can’t be upgraded to support Ethernet. The carrier will turn up this transparent LAN service later this year, Ballart says.

BellSouth bullish on MPLS

BellSouth is another MPLS proponent. The carrier announced plans to deploy terabit-capable core IP routers with MPLS to converge all its network traffic – including frame and ATM – onto one backbone.

MPLS will let BellSouth launch a new network-based VPN offering by the end this month. BellSouth is finalizing a deal on an edge MPLS router buildout for this service.

“The great part is that MPLS allows you to mix and match,” says Mark Kaish, vice president of data product management for BellSouth. “A small location could come in on DSL and that’s DSL at Layer 2, not Layer 3, so that removes the security concerns.”

Larger locations could use frame, ATM or even Ethernet, Kaish says. With MPLS at the core, the access technology wouldn’t matter.

BellSouth is devoting 12% to 15% of its revenue – or roughly $2.5 billion to $3.2 billion – to capital expenditures this year. That’s down from $3.8 billion last year and about half of the $6 billion spent in 2001.

In addition to MPLS-enabled edge and core routers, BellSouth is purchasing SONET multiservice provisioning platforms for edge and core applications, and upgrading its metropolitan Ethernet switches.

Multiservice gear will be deployed for customers subscribing to managed wavelength services that support IP, SONET, ATM or Ethernet for connecting storage networks, data centers, mainframes and high-bandwidth-demand locations.

BellSouth offers metropolitan Ethernet services in its region, at speeds of 10M, 100M and 1G bit/sec. But the carrier is upgrading its metropolitan Ethernet switches to offer better QoS and virtual LAN (VLAN) stacking, Kaish says. VLAN stacking would let a customer have multiple VLANs coming out of one port, which isn’t possible with the older metropolitan Ethernet equipment.

Qwest’s game plan

Like other RBOCs, Qwest is focused on bringing multiple services onto a unified network backbone. Unlike the others, the carrier isn’t entirely sold on MPLS.”At this point I can’t say MPLS would be the grand strategy,” says Augie Cruciotti, executive vice president at Qwest.

Qwest has turned up MPLS Fast Reroute for network recovery and is working on some MPLS-based services. But the carrier doesn’t see an immediate need to implement traffic engineering, a key application for steering traffic onto distinct routes for QoS and other deterministic capabilities.

Qwest says it will consider traffic engineering as traffic patterns change to ensure it’s a good business decision.

But Qwest needs a way to bridge its separate IP, ATM and frame relay networks. Customers currently cannot mix and match these technologies across different sites if they want the sites to be able to communicate.

“Once you leave the customer premises, we’d be going to a single backbone,” Cruciotti says. “We’d take the frame, ATM, IP, and map it all together.”

Qwest is attempting to get its frame and ATM services to work together. Cruciotti expects the project to run into next year.

Qwest is the least-aggressive RBOC when it comes to spending this year because the carrier finished its local and long-haul network buildouts last year. It also is distracted by investigations into its accounting practices by the Securities and Exchange Commission and U.S. Congress, and by its red ink – Qwest recorded a net loss of $35.9 billion in 2002, and a $4.8 billion loss in 2001.

So the company doesn’t plan on making large network capital expenditures this year, Cruciotti says. It will spend between $2 billion and $2.8 billion this year vs. $3 billion last year, and $8.5 billion in 2001.

“There’s no new grand technology,” he says. “We’re basically going with what we have in place now.”

For optical services, the carrier will incrementally deploy optical add/drop multiplexers, multiservice provisioning platforms and point-to-point dense wavelength division multiplexing. In Ethernet, Qwest is tracking the progress of the IEEE 802.3ah standard that will enable transport over copper. Qwest says this standard could drive demand for metropolitan Ethernet among its customers.

We can’t hear you now

The RBOCs are much less aggressive on next-generation voice. Though all are interested in offering end-to-end VoIP, they are still mainly kicking the tires of softswitches and media gateways in lab trials.

Momentum seems to be building gradually in IP Centrex. SBC rolled out IP and DSL Centrex services last year. Verizon is evaluating gateway and softswitch equipment in an IP Centrex trial in Chicago. BellSouth is working on IP-enabling its Centrex services for smaller businesses, and Qwest is looking to set up an IP Centrex hosting environment that would run an application emulating call control and call setup.

Qwest and BellSouth run private VoIP networks for some enterprise clients in which the carriers resell VoIP customer premises equipment and then provide WAN transport for that traffic. But quality is, has been and probably will continue to be an issue for VoIP, these service providers say.

“The quality we need isn’t there yet,” BellSouth’s Kaish says. “So we’re working directly with the vendors to come up with something that will deliver ‘five nines’ reliability.”

“QoS is still an issue,” says Teresa Taylor, executive vice president of Qwest product management and pricing.

Despite the challenges, RBOCs continue to spend billions of dollars. And as certain restrictions are lifted, the purse strings should loosen up a bit more, carriers and observers say.