EXCLUSIVE INTERVIEW - The future is integration, says Cisco

Jayshree Ullal is vice-president of Cisco's enterprise line of-coming trends. business, responsible for the full range of the network giant's campus Lan products.

Ullal joined Cisco through its acquisition of Crescendo. She took time out to give Network News an exclusive interview.

Network News: What do you see as the major trends in the networking industry this year?

Jayshree Ullal: Voice, video and data convergence. The catalyst is not just the PC. Mission-critical data is largely voice carried by PBXs and central office switches, but there will be a move in the enterprise to carry this traffic via IP.

There will also be a continuing scaling of performance in bandwidth from shared 10/100Mbps to Gigabit and Terabit speeds.

A big theme will be application-aware networking. The campus infrastructure has to respond and deliver increased speed. So we need congestion management.

Lastly, there will be more broadly-based end-to-end security with user authentication.

What is the most strategic product for Cisco's future?

While there's not a single strategic product, the Catalyst line of switches and the Cisco line of routers and its evolution to support voice, video and data integration is the most strategic.

Despite vendors talking up convergence, users remain cautious. Why is that?

The trend is there for voice, video and data networking. For the last 50 years we've had telephone circuit switching, but for cost reasons we'll now see a migration.

The technology is there, if not the actual manifestation of integration.

Customers are waiting for power and high resilience.

BT Syncordia, a key Cisco customer, ran a large trial of Voice over IP (VoIP) and found that the market will be driven by applications NOT cost benefits. Do you disagree?

There is no dearth of applications that will drive convergence. One application is call centres. Another is voicemail. Why should you have to pick up a phone to find out how many messages you have? Voicemail can be integrated to IP layers and written to a browser.

VoIP telephony will become increasingly important with the high-availability technology Cisco is able to deliver.

But reliability remains a key issue. People's data networks often crash, whereas voice systems are orders of magnitude more reliable. Why would they want to rip out a working PBX?

The attitude that 'if it's not broken it's not changed' won't persist.

In circuit switching new technology has not been an issue. But every time you add a PBX line you add $700/$1,000 per port. Then you have to overlay wide area costs.

We're catching a market in transition, not creating technology for the sake of it.

Now 75 per cent of traffic is voice. In three years' time it will be the reverse. Phone and internet applications will be unified.

The issue is high availability in data networks: how to take dial tone and make it web tone.

IP networks haven't been as reliable because they are changed more frequently.

The issue is how to build high availability for Terabit speed performance.

There will be convergence around IP and one common technology.

Will customers implement voice-data integration when they have other worries, such as Year 2000?

Economic forces and Y2K are definitely factors that always loom. Yet customers are deploying separate voice and separate data networks today.

The significant cost savings and ability to deploy new internet applications for data and telephony has already made customers deploy this. 1999 is the year of trials. 2000-2005 will be the years of major installations.

Moving on, what kind of technologies will you be developing?

We will extend DSL and cable modem delivery through partnerships. We won't just be developing technology for voice, video and data integration, but building a wireless infrastructure, through our alliance with Motorola.

How important is the delivery of Windows 2000 (NT5) for the networking industry?

NT5 is a critical element in delivering Active Directory.

Our directory strategy is working through the DEN (Directory Enabled Networking) initiative, embracing LDAP and existing directories. There will be further tighter integration between the OS and the infrastructure and that won't be platform-specific.

We believe in Active Directory but we're not going to 100 per cent standardise on Microsoft only. That's why we'll support LDAPv3. We're directory-neutral.

Tighter links with Microsoft will give us greater potential when NT5 comes - but this will not preclude links with other directories.

Supplying APIs to existing directories is critical, which is why we're providing interoperability with NDS.