Subscribe / Unsubscribe Enewsletters | Login | Register

Pencil Banner

Competing on multiple fronts helps Broadcom, CEO McGregor says

Stephen Lawson | Feb. 13, 2014
Broadcom got a jump on Mobile World Congress this week, announcing two steps forward in its fledgling LTE silicon business. On Monday, the company introduced a turnkey solution for LTE smartphones to be priced under US$300. On Tuesday, it announced a test, on a live carrier network in Finland, of a high-end handset chip that can use so-called Category 6 LTE with speeds as high as 300Mbps (bits per second).

The guys who are going to look at Arm servers are the guys who own their own applications. Because they don't care about x86, they care about DMIPS per dollar per watt. The Arm processors are fundamentally a better cost equation for that. They're more power efficient and they cost less. I think you're going to see [webscale companies] start to build whole data centers that are based on Arm architecture.

The early stage is that people have to get over the fact that it's a new instruction set and new architecture versus what they've had in the past, and these cloud guys will be the first guys to do it. The next challenge is that Arm processors, just by themselves, are commodities. That's a testament to how good the stuff coming out of Arm is. So I think you're going to see a commoditization of pure servers. I think what's interesting is, how do you start integrating a processor with, for example, switches and controllers and all the rest of the system to create an interesting architecture that's different from what we have today. And that starts getting interesting to a company like Broadcom.

The whole concept of NFV is very interesting here, where you morph the network to move to apps processors versus network processors. And that really plays to our strength. We've announced some new products and new architectures that are very interesting in the high-end network processor space.

IDGNS: SDN started as one specific idea and has become several different ideas, depending on who's talking about it. What role is Broadcom going to play in all this?

McGregor: Broadcom is delighted however this goes: We are agnostic on it. Most of the guys doing SDN are doing it with Broadcom chips. We put some things in our chips to help them. Back to my comment earlier about CIOs being the most risk-averse people on the planet, NFV is interesting because you can add NFV capabilities into an existing system incrementally. SDN, you either need a green field, or if you're adding new applications, you could add them in an SDN fashion on top of your existing framework. So, it's a little harder to just add SDN than it is to do NFV. We think SDN's very interesting. We think people will deploy it, maybe just not quite as fast as some people are hoping for.

IDGNS: Have the patent wars gotten out of control?

McGregor: Yes. I would greatly favor patent reform in this country. We have a problem now that is impeding innovation and impeding commerce. [One thing] I favor would be more proportionality: You should be able to extract value from a licensee to the extent that your patent contributes to the overall thing, but in proportion to all the other patents that do. For example, there must be a million patents that apply to a smartphone, but everybody who goes to court asking for royalties on a smartphone wants 5 percent. That math doesn't work. It has gotten out of control, and we'd like to see a more orderly market.


Previous Page  1  2 

Sign up for Computerworld eNewsletters.