The telephone systems’ intelligence lies in the switch. Though, data communications have added elements like remote access servers, DSL Access Multiplexers (DSLAM) and CMTS to this component of the network, it is the good old circuit-switched telecom switch that takes a bulk of the traffic hitting the network. What Cisco is saying is that the central office switch solution in use today is inflexible and proprietary, forcing service providers to source all the hardware, software and applications from a single vendor. This "big iron"—as it calls it, must give way to open programmable switching products that are based on open call model and standard API and supports cost-effective and rapid development of core transport and enhanced services like calling cards. These products would have applications residing on a host computer that controls the switch through a command and report relationship. It is this segregation of the application from the call control and platform elements that is fundamental to Cisco’s strategy. But all said and done, the postscript from Cisco is that open programmable switching platforms can work adjacent to older proprietary switches, providing core transport for IP traffic or serve as a mixed environment gateway/node for packet-based networks.

Today, while there is a consensus on an ultimate cross over to packet switching, many telecom equipment vendors don’t see circuit switches disappearing in the near future. Also, vendors don’t agree with Cisco that IP will necessarily be the single standard, they see a bigger role for the other packet (actually cell-based) technology—ATM—in the short and long term.

Also, circuit switch majors like Lucent, Nortel, Ericsson, Siemens and Alcatel haven’t approached the open API development concept on one voice. While some like Lucent have willingly crossed over to the open API camp, others have stuck to adding the new applications to their "big iron" on their own. And again, in this space there is a breed of companies like Sonus Networks, Unisphere and Nortel that have developed a new class of purpose-built gateways supporting IP, ATM, and TDM-based traffic.

The Forecast

Of late, there have been some wake-up calls from stock analysts that the capex of telecom service providers might flatten out or even decline in the year 2001, which immediately had a southward effect on major telecom vendors’ stocks. Not even Juniper and Cisco were spared. Experts, however, feel that recent announcements of capex reduction by major service providers could be routine, considering the volatility of telecom capex. They also point out that the services which today’s carriers want to sell to businesses—Virtual Private Networks (VPNs), guaranteed high-bandwidth and Application Services—cannot be deployed over the old voice networks. Service providers run the risk of falling behind competition, if they do not invest in a new network. The total telecom spending for this year, estimated at around $300 billion, is expected to grow at a rate of about 18 percent—touching $350 billion next year.

According to RHK Inc., $23.6 billion is expected to be spent on optical networking. In this market, Cisco holds just 3 percent marketshare, while Nortel (45 percent) and Lucent (25 percent) together hold 70 percent of the market. Bulk of the revenues for Nortel and Lucent come from selling SONET/SDH equipment, ATM/TDM-bases multiservices, WAN switches and long-haul DWDM equipment. It is in the metro-DWDM and SDH/SONET markets that Cisco has mainly placed its hopes. Analysts believe the SONET/SDH market is entering a time of change. It is this trend that Cisco would want to encash with its line of gigabit switches—IP as well as multi-service—and optical routers.

In the core switching and routing market, Cisco is much better placed. According to Cahners In-Stat Group, the ATM switch market is expected to touch $4 billion in 2000. Lucent has a marketshare of 40.2 percent, Nortel 20.7 percent, while Cisco came third with 17.1 percent. Here again, Cisco sees its share increasing as ATM gets replaced by IP. But, In-Stat analysts feel that it will be at least 18 months before alternative technologies threaten to displace ATM switch sales.

Actually, it is in the familiar territory of routers that Cisco is feeling the pressure—from a company that allegedly took away its cream of engineers working on routing. Juniper has been hitting headlines more than occasionally with its "faster than Cisco" routers and it has, unlike many start-ups, figures to prove its impressive aggression in the core router space. In an estimated $11 billion overall router market for 2000, Juniper’s performance is expected to increase its marketshare to about 4 percent. This leaves Cisco with the worry of hitting below the 80 percent marketshare in 2001, something that has not happened for a long time.

While Cisco core routers may be hit by Juniper’s terrabit routers, in the edge router scene, Cisco is expected to hold on to its 80 percent plus marketshare for at least a year, as Juniper is only starting to address the market.

It is in the telecommunications switching and access space that Cisco has to really market hard to make an impact. Though data communications equipment is being deployed by ISPs and carriers, they have accomplished just a few cases of replacing a circuit switch. In the datacom access market, Cisco’s performance has been quite excellent. It has a race for leadership with Lucent in the dial access server/concentrator market. In the other developing access markets like DSLAM, cable and broadband aggregration, it is in the top 3 positions. However, the access space is a small market at approximately $8 billion for the year 2000. The access market contributed only $2.8 billion to its $18.9 billion revenue for fiscal 2000, i.e. about 25 percent marketshare.

Contrast this to $48 billion that telephone companies are expected to spend on circuit switches according to JP Morgan, and one sees a huge gap. Telephone companies are expected to spend seven times less than that on data internetworking devices.

No marks for guessing—this is the space where Cisco has sets its eyes. With packet technologies, it hopes to make its access equipment redundant and reliable enough to handle millions of calls every day. Then its equipment would have made the good books of the QoS-fussy telecom engineers, thereby enabling it to replace today’s circuit switches. Until then, it has to be satisfied with its share of the packet access market put at $15 billion by 2004.