PDF of this tutorialArchitecture Limitations
While the stage has been set for a new competitive landscape, the existing network architecture has prevented its full realization. In many markets where competitive local exchange carriers (CLECs) are challenging incumbent local exchange carriers (ILECs), services are merely resold by means of the existing infrastructure. In fact, industry experts estimate that more than 90 percent of all CLEC access lines are passively resold. Facility-based CLECs attempting to build new networks have been hampered by the high cost of large, proprietary Class-5 switches. With a typical up-front cost of $3 million per deployment, traditional switch technology has prevented CLECs and other next-generation service providers from significantly altering the cost models long employed by the ILECs.
The existing proprietary Class-5 switching infrastructure has also hampered service providers' ability to create and deploy innovative new services to gain a competitive advantage. The proprietary architectures of the existing switch platforms have made network operators entirely dependent on switch vendors for new software applications and upgrades. These software upgrades and new services often take years to develop, are extremely expensive, and remain the exclusive property of the switch manufacturer. This makes it difficult for service providers to differentiate themselves and limits the services available to consumers.
Network Transformation
Today, however, the U.S. telecommunications market is poised to enter a new era of competition—an era that realizes what the Telecommunications Act of 1996 promised but did not deliver. The PSTN will be transformed as voice, data, and enhanced services converge. A new network architecture is about to emerge, driven by new technologies, new demand, and increased competition.
Today's network is divided into two elements: the PSTN and the PSDN (see Figure 1). The PSTN consists of large, centralized, proprietary Class-5 switches with remote switching modules (RSMs) and digital loop carriers (DLCs). This architectural configuration, representing $250 billion in network investment and hundreds of billions of dollars in annual service revenues, has changed little over the last several decades.
Figure 1. The PSTN and PSDN
In contrast, the substantially smaller PSDN—consisting of network points of presence (PoPs) and remote access devices—is growing at a dramatic rate. The growth of the PSDN is driven by the Internet, intranets, virtual private networks (VPNs), and remote access. However, the PSTN continues to be the principal means of delivering data services. According to Dataquest, 46.5 million analog modems will be sold in the year 2000; nearly all personal computers (PCs) purchased today come equipped with a 56–kbps modem.
Many industry pundits claim that packet-switched voice will displace circuit-switched voice in a matter of a few years. Despite the hype, however, voice over Internet protocol (VoIP) has yet to establish itself to any significant degree. Furthermore, if packetized voice is actually to displace circuit-switched voice, a decade or more will be necessary. In fact, the March 1999 edition of the Pulver Report estimates that in the current $400 billion global minutes marketplace, 1999 revenues for IP–telephony minutes will be less than $150 million.
Some form of convergence between the PSTN and the emerging PSDN is inevitable. The trouble is, no one is precisely certain about how and when this convergence will take place. Nevertheless, two points seem clear:
- With the existing CO switch infrastructure representing a multibillion-dollar investment, it is unlikely operators will opt for a wholesale replacement any time soon.
- While packet-switched technologies like IP and asynchronous transfer mode (ATM) will dominate in transport applications and are well-suited for Class-4/toll circuit switching, Class-5 time division multiplexing (TDM) will be required in the network for some time. Class-5 switching will become more distributed, reside at the edge of the network, and become integrated with enhanced service offerings.
In other words, circuit switching and packet switching will coexist for quite a while, with IP, ATM, and TDM all playing complementary roles.
