As vice president of research and development at BroadcomDr Ed Frank has a wide range of technologies to oversee.
The company’s chips and software solutions are embedded in cable and DSL modems and digital set-top boxes in the homedigital televisionshigh definition DVD playersnetworking equipment in the enterprisewireless-enabled laptop and desktop computersand advanced PDAs and cellular phonesamong wired and wireless equipment.
Broadcom is a collection of companies that have been acquired over the yearssuch as Element14 in Cambridge and Bristol and AlphaMosaic in Cambridgeeach with its own engineering.
Dr Ed Frank joined Broadcom as Senior Director of Engineering of Home & Wireless Networking in May 1999 following Broadcom’s acquisition of high speed home networking chip maker Epigramand was elected Vice President of Research & Development in 2003. Dr Frank was a co-founder of Epigramhaving been a co-founder and Vice President of Engineering of NeTpowera computer workstation manufacturer and at Sun Microsystems. He is a named inventor in 37 US patents and is a Trustee of Carnegie Mellon University.
NF: What can you do as vice president of R&D in this situation?
Ed Frank: I spend my time in three areas – new marketscustomers and engineering internallydistributed around the business lines. We have done a very good job at avoiding the fiefdoms. With new markets we need to those markets that are very high volume. For example I spent a lot of time last year going to the mobile multimedia player market. To be successful we really need to find these markets where with a US$30m to US$40m investment in chip design there's enough market to give us a returnso you're looking at finding a billion dollar market.”
NF: Broadcom is expanding its design centre in Bristol to build up the set top box design capability alongside the DSL modems chip design already here. What is the trend you see there?
Ed Frank: The challenge we have is on multiple tuners on a chipwhich is less a technology one than a market one. The issue is getting a set of projects to align so we can do the integration or whether to evolve the two parts separately. It's been better to move to the next process node quickly then have things integrated.
NF: Broadcom now also has a strong focus on wireless as a new businesswith a single chip cell phone design and moves into the 3G space with an HSDPA chip based on technology it acquired with Zyray and a strong position in 802.11g WiFi and now looking at 802.11n. Do these technologies come together and do you have to wait for standards to solidify?
Ed Frank: I think standard technologies generally always win and that’s how Broadcom has become what it is today. We have silicon today that supports the draft standard of 802.11n and our approach is to be fully software upgradeable to deal with any changes that may happen. Even for non-processor-based devices Broadcom has been veryvery flexible and adaptableand we have been doing that for ten years. I think we will see multiple wireless technologies and basebands in a single chip as the processes are maturing.
NF: Broadcom has deliberately focussed very heavily on standard CMOStaking advantage of the process developments to push more integration and lower costs to drive the business. Where does the technology go now with 65nm CMOS coming now and 45nm processes on the horizon?
Ed Frank: We are having to invest in both continuing aspects of the current generation and more aggressively for the next generation – there’s more overlap than we used to see and more tricky to plan for products in a particular process node.
But the reality is that we are going to be the heir apparent and the old guard just can’t keep uplike we did with 802.11. We are hyper aggressive because we will eat our young– we are much more willing to come out with new silicon that replaces our existing silicon.
We are not perfect here but we aim wherever possible to be first to market with the next generation – we did it with gigabit Ethernet and with 802.11g. Part of Broadcom’s added value is Broadcom – we are dealing with people who need to have confidence that we will be a supplier for the next five to 10 years. We aren’t going away and we continue to grow faster then other billion dollar semiconductor businesses around and even faster then a lot of smaller companies. We are still in acquisition mode."