Teleglobe Communications Corp. (Reston, VA), the world's third largest long-distance carrier, has announced deployment of a state-of-the-art Internet protocol/Asynchronous Transfer Mode (IP/ATM) public network over fiber that will integrate Internet services, voice, data, and video traffic over a single network interconnecting the world's 160 largest cities over the next five years. Known as GlobeSystem, the $5-billion network will feature advanced optical switching, protection, and routing technology. Teleglobe Communications hopes to ride GlobeSystem to the top international spot in IP networking by offering what the company describes as the world's first high-reliability voice and data packetized network with customer-specified quality-of-service parameters.
"Teleglobe is taking a greenfields approach with an IP public-networking platform," says Jeanne Schaaf, a senior analyst for public networks at Forrester Research. "It's an [example of an] international carrier branching out to get a foothold [in the global broadband Internet]. Teleglobe is based in the U.S., and because the U.S. is centric to the development of the Internet, the GlobeSystem network will capture a lot of Internet traffic on a global basis, as more and more countries come into the adoption curve."
From a business perspective, the Teleglobe network will address a need among carriers, corporate customers, Internet service providers, and content distributors to find a seamless, optically protected global network to transport data, voice, and Internet video regardless of local-access method or air interface (e.g., IP, ATM, TDMA, and CDMA). "Customers can have dedicated lines into our 160 common switching platforms [known as 'GlobeCities'], and we'll be able to take their services and transport them around the world," explains Andrew Burroughs, vice president of global marketing and product management at Teleglobe." A total of 47 GlobeCity platforms will be operational within the next 12 months.
Because of its global communications presence, "Teleglobe is best positioned in the international marketplace to serve data and Internet needs [internationally]," says Daniel J. Fletcher, president of Lehman Bros. (New York City). "Teleglobe's long-distance network is third largest [behind AT&T/BT and MCI/WorldCom] in terms of countries and reach; it has network services in parts of the world where data and Internet growth is expected to be significant."
Deployment plans for the network are aggressive over the five-year horizon. Teleglobe claims it will use 160-channel dense wavelength-division multiplexing (DWDM) to boost network capacity in the Northeast (including New York, Toronto, and Montreal) and has added new fiber linking the northwest United States and Canada. The company also plans deployments of next-generation fiber (again using 160-wavelength DWDM) in an intra-European ring linking London, Amsterdam, Rotterdam, Brussels, Antwerp, and Paris, followed by a second European ring linking Germany and several other countries. Further, the company is upgrading its North American Internet backbone to OC-192 (10 Gbits/sec) by the fourth quarter of this year, boosting customer Internet networking speeds.
"Within five years, our network will increase to 180 times its current capacity, from 1 million DS-3 fiber miles to 124 million DS-3 fiber miles," reports Teleglobe's Burroughs. "The network will be based on advanced IP and ATM platforms. Anyone can interconnect to our technology regardless of their own access method and network protocol."
The company claims it will be the first to apply optical technology that fuses wavelengths directly onto routers and ATM platforms. Customers will then be allowed to define their own service parameters, giving the global IP network an ATM-like quality-of-service flavor. Further details are sketchy, however. According to Steve Heap, vice president of global network planning and engineering at Teleglobe, optical switching/protection systems will eventually become integral to each GlobeSystem node. So will optical add/drop multiplexing. But photonic switching and restoration are still planned further down the five-year curve.
"By using optical switching," says Heap, "we will remove a costly electronic layer of the network [associated with SONET and SDH, which currently performs problem detection and alternate routing]. Optical switching and protection are cheaper and more reliable. As far as fusing data to routers, the principle is this: Data comes out of the router at the wavelength level-we're removing the electronic layer-and goes into an optical switch and into the fiber. Some of the traffic will come directly from the customer's router to our backbone router, then to the optical switch, then to the DWDM, and [finally] into the physical fiber."
Heap says both Nortel and Lucent Technologies are working on the optical switching/protection equipment components of the new network, but Teleglobe has not yet announced final equipment selections for most of the GlobeSystem project. Alcatel, however, has been awarded a contract to supply the Optinex 1640 DWDM technology for three early terrestrial projects (two in Europe, one in the United States), part of GlobeSystem's Phase I build-out. The Optinex DWDM system will multiplex wavelengths over 1700 km of dark fiber on the new Teleglobe European fiber-optic ring interconnecting cities such as London, Amsterdam, Brussels, and Paris. Alcatel's technology makes GlobeSystem the world's first 400-Gbit/sec optical network designed to operate over both terrestrial and submarine fiber-optic cable. It will deliver higher capacity per fiber than any existing European network.
"Our initial [Teleglobe] project is probably worth between $10 and $20 million," estimates Tim Krause, associate vice president of product management, Alcatel optical backbone networks. "The Optinex 1640 architecture supports up to 240 wavelengths, and the Teleglobe initial deployment will be based on 40 wavelengths [each transmitting 10-Gbit/sec-OC-192 capacity], which can be expanded up to 160."
Alcatel is also in discussions with Teleglobe regarding deployment of "repeaterless" submarine cable systems such as those used on large lakes, rivers, or channels, up to 200 km. "We've done business before in the area of submarine-plant optical networks," Krause says. "Only our 1640 DWDM is able to achieve 200-km links when engineered for 40 channels at 10 Gbits each [e.g., OC-192]."
The longer-term prospects for GlobeSystem are even more complex. Teleglobe is planning a second phase of its network beginning in 2001 that will include deployment of a transatlantic fiber-optic network system that uses as yet unspecified "next-generation" technology. In 1994, Teleglobe was the first to deploy a transoceanic fiber cable system based on Synchronous Digital Hierarchy (SDH) technology.
"This will be our fifth cable across the Atlantic," says Teleglobe's Burroughs. "While we're not going to build every cable in the world, there are certain routes that Teleglobe needs to have facility independence. We've announced a cable going from Pennant Point Canada to land in the Netherlands. And we're looking at laying a festoon cable [that goes around the Asian subcontinent]."
Because of booming Internet traffic, Teleglobe is not concerned about a fiber "glut"; "There are certain areas of your corporation where you want to have control over your destiny," Burroughs says.
Analysts appear bullish about Teleglobe's prospects. Forrester's Schaaf argues the company can get a leg up on Qwest and Global Crossing by announcing and deploying premium public IP first. Lehman Bros.' Fletcher agrees: "There are two general categories of 'why' Teleglobe is doing this. First, they have the expertise and backbone systems supporting it; they know the local governments and regulations, and it's less of a risk for them and some of their competitive interests. Second, there's something to be said for executing first. A number of companies say they're going to build a [global Internet voice/data network] and haven't yet. On the cost side, because Teleglobe is already a large network, the incremental cost is not as significant as building out an entirely new network. This gives them a cost advantage decision." q
Arielle Emmett has written exclusively about a wide range of communications issues. She is editorial director of Lightwave's sister publication, Wireless Integration.