Avici Systems - broadening market focus under new CEO
February 28, 2005		Email this page

Q&A with Dr William Leighton, CEO, Avici Systems.

Bill Leighton has over 25 years of senior management experience spanning research, engineering and business units in telecommunications. Prior to joining Avici Systems, Bill worked at AT&T in a variety of roles over the last 28 years, most recently as Vice President of Research at AT&T Labs where he led the team responsible for AT&T's core and directed research program. Prior to that, he was the Vice President of Technology and Services Strategy at AT&T Labs. Specifically, Bill led the teams that designed and built AT&T's Frame Relay, ATM, VoIP and IP networks.

Introduction

Avici Systems, headquartered in North Billerica, Massachusetts, was founded in 1996. The company has developed a range of high-capacity, carrier-class core routing solutions, which are marketed to service provider customers worldwide.

In January 2004, Avici formed a strategic relationship with Nortel Networks, whereby Nortel offers Avici products alongside its portfolio of networking solutions. Avici has systems deployed with carriers including AT&T, China Telecom, France Telecom, IP-Only, Telecom Italia Labs, and SURFnet.

Could you give an overview of the steps you have taken since assuming the role of CEO?

When I was appointed interim CEO with Avici in November 2004, having been a member of the board of directors since May 2004, it was obvious what the fundamental problems were - customers and revenue. During my first week with the company, one of the first things I did was to cancel all staff meetings and devoted this time instead to a strategy planning session with the extended senior management team, primarily to pinpoint the strategic basis going forward.

As a part of this strategy session we conducted a root-cause analysis of every contract win, loss and no decision. I had every member of my management team take a different customer win or loss scenario.

This analysis revealed some interesting patterns, the first being that Avici has spent the past two years chasing features - with repeated statements along the lines of 'we lost this contract because we did not offer feature x, and this one because we could not provide feature y'. But if you look deeper, you find this wasn't the real reason; it's just what the customer told you to get rid of you. From the customer's perspective this is an easy and effective way to eliminate a vendor they want to rule out. We went through it to ask: Was that really the reason? If you had done that, or if you had feature x or y would you have won it?

When you're in a sales situation, frequently you only hear what you want to hear. Customers have learned how to deal with vendors in these multi-vendor environments. They have their own language. Having been the customer, I know the language. If a carrier tells you, you were eliminated because you didn't have feature x, typically that really means, "I'm going to find a clean reason you can't escalate around".

In reality, is it because you don't have feature x? If I'm a carrier and you're a vendor and I want to just get rid of you cleanly, I just say, "Well, you don't have feature x, and I can always find a feature x". By telling you this, you don't have recourse and you can't escalate it.

The team analysed the data generated from this exercise to form a better view of the carriers' needs, and thus a clearer picture of Avici's market.

Using the knowledge gained from this session, senior management devised a three-pronged strategy to take the company forward, starting with the near-term goal of addressing tactical execution - focusing on how to win business and pre-empt the competition. We are having discussions with carriers focused on what their real problems are as opposed to what features they are looking at and have also identified new opportunities that don't restrict Avici to, for example, only tier-one carriers. We are broadening the market focus and also revising what is assessed to be an opportunity.

The primary objective in the medium-term is to develop product differentiation strategies. Originally Avici offered a very strong product that was technically superior to comparable systems from Cisco and Juniper - this is why AT&T signed up as a customer.

At that time the GSR system from Cisco had 8 slots and the Avici TSR product had 40 I/O slots; the company claimed that it was possible to link fourteen TSRs together as a single router. The solution was highly scalable and at the time, AT&T believed they could link at least two systems together for 80 I/O slots, which was an order of magnitude higher than what they could do with Cisco.

Avici can still claim an advantage in the market in terms of system scalability, plus a significant lead with regards to reliability. The task now is to extend the degree of competitive advantage enjoyed four years ago.

With regards to product differentiation, an area of interest is looking at what functions can be shifted from the edge to the core of the network. IP networks have now moved into a space occupied by voice networks ten to fifteen years ago. In these networks, similar to what happened in voice networks, adding a new service generally entails installing another box, which adds complexity. Some of the functions performed by these boxes, spread around the network edge or the enterprise, could be carried out more efficiently by the core router platform. Network security is one example of a function that can more effectively be performed at the core of the network.

In the long term, the question centres on the fundamental approach to networking. I believe that networks are becoming far too complicated and complex, the goal has to be what I term simple, secure networks. Complexity breeds a number of problems that have yet to be addressed, not least of which is cost for the service provider at a time when they are under tremendous pressure with regards to operational expenditures.

Have you instigated a time-line by which Avici must achieve certain goals?

Reaching the near-term objectives will serve to validate the changes that are being implemented within the company. I cannot claim yet that the measures taken so far have translated into customer wins, but there are encouraging signs that the company is moving in the right direction.

Do you feel that Avici has an image problem, and that there is a risk it will come to be regarded as an also-ran, given that the company has been active for some time but has failed to achieve more than moderate success in the market?

I would certainly accept that Avici has an image problem - one that its competitors are happy to exploit. However, I believe this is due as much to external factors beyond the company's control as to mistakes made by the company or any inherent weakness in its products.

When Avici launched its core router product around 2000, AT&T was the only major carrier that had not recently upgraded to a Cisco or Juniper platform, and the carrier chose Avici's product.

During this period Juniper successfully built up a significant customer base to become the second source for core router systems behind Cisco. Due to the cyclical nature of this market the carriers then stopped investing in core routers, simply because they had recently completed overhauling their networks. This would have been the case irrespective of the subsequent downturn in the industry. The main demand during such a period is for line-cards to fill out empty slots on deployed core equipment - predominantly Cisco and Juniper systems. The cycle is once again reaching the stage where carriers are beginning to upgrade their core networks.

During the lull in the market, Avici developed its relationship with AT&T, got the product to its current state, set up channels to go to market, and won some smaller contracts, including SURFnet and recently IP-Only Telecommunications in Europe.

On the positive side, Avici is not an unknown player in the industry, moreover its products are still generally considered to be technically superior to other core router solutions on the market. Doubts centre on why the company has not achieved more success - it is primarily a matter of confidence. I believe that with proper focus and execution Avici can break out of this vicious circle.

Assisting this transition will be the fact that Avici operates within a very tight community where everyone knows what everyone else is doing, and where relatively minor events can have a noticeable effect.

For example, AT&T's announcement of a 'five-nines' SLA for its IP network caught the attention of the carriers. AT&T is also the only major carrier that has not needed to do an overhaul of its IP network - they don't need to because they've been able to add in-service capacity as needed without forklift upgrades, and their IP network is operating at a telco-grade reliability level.

This is prompting carriers to reassess the equipment they are operating and how it will enable them to develop their position in the market. I am aware of carriers that have found they simply cannot currently afford to expand their Cisco or Juniper-based networks to meet their business objectives.

Was the recent IP-Only contract the first to come as a result of the partnership with Nortel?

No, the first contract to come out of the Nortel partnership was with SURFnet 76 days after we announced the relationship.

How did the IP-Only win come about - was the company an existing Nortel customer?

IP-Only has some installed Nortel DWDM equipment, but had a specific requirement relating to its IP network. The company was operating a competitor's systems in the core of the network and was looking for a solution, whether from Cisco, Juniper or Avici, that would meet a number of criteria.

For IP-Only, the main attraction with Avici's solution was the common features, line-cards and software across the product range - spanning the 44-slot TSR, 22-slot SSR and 10-slot QSR platforms. The carrier realised that it could have access to the same features and technology as AT&T, which is running a greater than 'five-nines' network carrying over two petabytes of data per day, in a platform scaled to its needs. Avici's solution allows IP-Only to deliver high-reliability services as well as advanced services reliant upon the non-stop routing capability of the product.

Did IP-Only state why it had opted for Avici's equipment as opposed to that of the competitor with systems already installed?

IP-Only is a carrier experiencing fast growth in its business, with some very demanding enterprise customers and as well peering relationships with other carriers. The company therefore has a need for the core router technology and capabilities of the Avici QSR.

Cisco, for example, is claiming scaling capabilities, but only in its CRS system, which is simply too large - and too expensive, and it is the same story with Juniper's T640 product. Avici is the only vendor to combine the technology and capabilities that IP-Only required in a footprint suited to their needs.

How long was the sales cycle with IP-Only, from initial contact to contract award?

With IP-Only the process took less than six months, although with larger carriers it would typically be more than one year.

For a carrier of IP-Only's size, would six months be typical?

Yes. In the case of IP-Only Avici was dealing primarily with two people - the CEO and the CTO. They knew what they wanted and were aware of what solutions were available and so were able to quickly discount rival offerings as impractical for their needs. The fact that there was no large committee making the decision also helped to speed up the process.

Is the lengthier sales process with large carriers simply down to the complexity of the networks, or due more to bureaucracy?

It is due to a number of issues but principally the fact that such companies begin planning earlier, and will devote considerable time to identifying and testing all of the technologies available on the market. At the end of the day, a major carrier is going to commit a large sum of money, and so wants to be certain that it makes the right decision. AT&T deliberated for around eighteen months, including testing from initial contact, before it deployed the TSR system from Avici.

Regarding Avici's relationship with Nortel, how many staff, excluding the sales force, does Nortel dedicate to the partnership?

There is an executive champion within Nortel who is responsible for the business logistics side of the relationship, plus product management, customer support, engineering, and of course sales. But for the most part activities relating to the partnership are fully integrated into Nortel's day-to-day organisation.

What level of resources has Nortel devoted to its partnership with Avici?

Nortel has directed significant resources to the relationship, as one example, when the agreement was initiated more than 100 system engineers and customer services staff were trained on Avici's products. If you add product management, marketing and sales the number is over 400.

Beyond the sales part of the relationship, there is also joint product development and integration, which means that Nortel and Avici commit resources in terms of engineering and product management activities. The agreement goes further than Nortel simply reselling Avici's products.

Does Nortel have a revenue target that it must reach to count the partnership a success?

Yes!

Do Avici and Nortel divide up the market?

Basically no, the market is not divided up between the companies along geographic lines, but rather where relationships exist. There are certain accounts with which Nortel is well positioned that they would tend to deal with, but Avici pursues any customer opportunities that we identify wherever they are.

What does the title 'preferred partner' signify in Nortel's relationship with Avici - does Nortel have an alternative supplier for core routers?

Preferred partner means that where Nortel has a need for a core router system, Avici is its partner of choice. To put it another way, Nortel does not have a relationship with another company offering a core router system.

How much involvement does Avici have in contracts delivered by Nortel - does Nortel perform most of the sales process?

Avici is directly involved with customers, whether or not the contract was won through the Nortel partnership. For example, Avici staff spent one week with IP-Only training their technical staff on the product. IP-Only commented on the fact that Avici was extremely responsive to its needs as a customer - from training, to support and system installation. It is our product, our name, therefore our reputation - we want to make certain that customers are happy with the Avici product and our responsiveness to their needs.

Does the partnership work in the opposite direction, with Avici bringing business to Nortel?

Yes. There have been two occasions so far when Avici has brought business to Nortel.

Avici has direct sales staff on the ground - it is not solely reliant on Nortel for its business; in some cases these sales staff identify opportunities that Nortel has not addressed. Where a customer requires a complete network solution, it makes sense to bid jointly for the contract.

In such circumstances, if Nortel wins business but Avici does not, is there any compensation scheme?

Avici does not earn commission from Nortel, unless Nortel sells our products.

With regards to the partnership it should be noted that unlike some reseller arrangements, the Avici routers are fully integrated as part of Nortel's data portfolio. What I mean is that to a Nortel sales person, Avici products are simply part of Nortel's portfolio - there is no disincentive to selling Avici equipment for the Nortel sales staff. In some reseller relationships, it is sometimes the case that the compensation structure discourages sales staff from actively selling OEM products.

Would you regard it as a problem if, in the future, the bulk of Avici's sales were generated via the Nortel relationship?

No. If eventually the Nortel channel, which is now beginning to come on line, were to be Avici's main source of business, I would consider it a very good thing. However, we are very focused on identifying and closing new business whether that is direct, through Nortel or other means. As I said before, our near term focus is to aggressively pursue new opportunities and close them.

Do you believe that this situation could lead to Nortel dictating how Avici should develop its products?

I don't believe there is a serious risk of this happening, principally because the expertise in IP core routing, which underpins the products, resides with Avici. In addition to this, from Nortel's perspective one of the benefits of the relationship is that as a small company Avici can achieve results fast!

This article is the copyright of Optical Keyhole. It may be freely distributed by any means in an unaltered form.