IDG Enterprise CEO Interview Series: Meru President and CEO Ihab Abu-Hakima

Meru Networks claims to have been the first company to deliver an 802.11n access point and is now riding that technology’s popularity as enterprises move increasingly to high-speed wireless networks. In this installment of the IDG Enterprise CEO Interview Series, Meru President and CEO Ihab Abu-Hakima speaks with IDGE Chief Content Officer John Gallant about what sets Meru apart from bigger competitors with broader networking product lines such as Cisco and HP, as well as what needs to be done by enterprises to manage and secure networks being flooded by iPads, smartphones, and other devices.

What makes Meru unique?

Meru was founded based on a vision that sooner or later most enterprises will operate day to day in an all-wireless environment. For us, a wireless environment is a wired data center, a wired backbone but all wireless everything else, all wireless edge. As we looked forward to what an all-wireless enterprise would look like, we said it would have thousands or tens of thousands of devices, many of them operated by humans, others machine to machine, but they would all be mobile or stationary but working in a wireless environment. And we built an infrastructure from the ground up to support density and mobile or stationary voice, video or data applications. We built this with the end-user in mind. We wanted the end-user to have an interactive experience accessing the content that they needed or the applications that they needed to get their job done.

To go into a little bit more depth about that, talk about the virtualized wireless architecture. What do you mean and how is that different from the approach that your competitors take?

If you look at the architectures from competitors, like a Cisco, Aruba, or HP, they’re what’s called micro-cell. By micro-cell, each access point has to be placed on a different channel and you go through an exercise where you have to place the access points far enough away from each other that they don’t interfere but not too far away that you have a gap in coverage. So it’s an optimization exercise that you go through and you design a network to support a certain number of devices and density. As you add more devices to that type of a network, you need to add more capacity. To add more capacity, you have to add more access points. If you add more access points, you have to redesign a network because you have to bring access points closer together and shrink cell sizes so they stop interfering. In every case, adjacent access points are on different channels. But as you do that, you are introducing more contention for the same bandwidth between the access points.

With Meru, what we do is the architecture has been developed so that every access point can be on the same channel as its neighbor. So you can layer a whole floor or a whole building or a whole enterprise with a single channel, or all of the access points running on the same channel. You can run each access point at maximum power because we’ve mitigated the interference. And we do that by coordinating among the different access points.

Now if you think about it, we’ve only used one channel in the 2.4GHz and the 5GHz bands and you have all of these other channels. So we can layer on another set of radios on the next channel and so on. We end up stacking all of the channels just like you would stack a wired Internet switch to handle more capacity. By doing this, we’ve essentially captured every megahertz that’s available in the 2.4GHz and the 5GHz gig bands for the enterprise and we’ve created this virtual pool of all of the spectrum that’s available. And now we can start partitioning this pool, megahertz by megahertz, depending on the service level requirements of the application running on the device. If it’s an HD video and you need 5Mbps-sustained throughput, that’s what you’re going to get.

Is that an approach that your competitors are likely to catch up with or are they taking a fundamentally different path toward the same end?

Well the competitors have all taken the same path, kind of an old path that’s been around for many years from the days of the first standard on access points. And then when controller architectures came out, that didn’t change, so all of the competitors operate the same way. They put adjacent access points on different channels and they try and deal with interference by shrinking the cell size by dialing down the power levels. That becomes problematic the more access points you bring into the network. So will they one day flip to the single-channel architecture? We don’t know.

You’re competing against some pretty significant companies like Cisco and Aruba. What’s the competitive landscape like right now? Where do you win business? Where do you tend to do better than them?

In every opportunity, we compete against Cisco and that’s not a surprise. Just about everybody is a Cisco customer for some product or another. And Cisco has done a phenomenal job over the last dozen or so years, seeding the market with wireless LANs.

As you add more devices to the network and as you start to run more converged applications (as you start to add the video and voice), each one has different data characteristics, and the needs of the enterprise move toward the sweet spot of Meru’s architecture.

The majority of Meru’s customers are Cisco customers who remove Cisco wireless and replace it with Meru wireless and continue to run Cisco’s voice over IP system. We plug right into existing Cisco switches. Though over a period of time, customers realize that the traffic pattern shifts and in the past, maybe 5 percent of their day-to-day network traffic went over the wireless LAN over a period of time, that’s now 50 to 80 percent of the traffic going on the Meru wireless infrastructure and more and more of their wired ports are no longer used. So they decommission those switches.

Depending on the vertical market, we see Motorola, HP, and Aruba as well. We’re smaller, but growing.

Talk to me about the concept that you mentioned at the beginning: the all-wireless enterprise. How close are we to it and what do you think are the remaining obstacles?

If you think about the architecture that Meru introduced, it’s fundamentally different than anything else out there. Invariably when we would go into an opportunity and talk to the IT organization, the reaction is typically: Well, we didn’t think wireless works this way and this sounds too good to be true that I can just pick up access points every 30, 40, 50 feet and put them on the same channel and go.

So we’ve tended to attract the early adopters and the lighthouse accounts, those that have been running into these problems ahead of others, that are trying to drive the use of more mobile devices whether it’s for data capture or wireless voice over IP and so on. It is not unusual now to find our customers running 90 percent of their day-to-day traffic in an all-wireless environment and for their wiring closets to be empty of wired switches.

How far away are we from this becoming mainstream?

Probably three or four years ago we felt that there needed to be some driving forces, one of them psychological, which we felt would be the 802.11n standard. Maybe more networking IT individuals would feel maybe with 11n that they would have more bandwidth, so could run more applications. So that psychological shift happened. And then the second one was most people in the industry — including myself — thought that the advent of smartphones or smartphones tipping over a critical threshold would accelerate this market. Smartphones did help, but the iPad really was the trigger point. So I would say we’re somewhere between five and eight years for this to be mainstream.

You mention the iPad. What aren’t enterprise IT folks thinking about with this flood of new mobile devices? What will be the impact and what will that do to existing wireless networks?

There’s probably an underestimation in terms of the impact of the number of devices that will appear on the wireless network, slowing down or in some cases making the wireless network inaccessible. It’s not unusual to hear about conferences or seminars where you have 50 or 100 people in a room and you may see five bars on your laptop but you can’t get on the network because a handful of devices have essentially hogged the whole bandwidth. We call it Wi-Fi meltdown — we’ve seen this more at conferences, but I think more enterprises are going to start to experience this.

And the second driver is the convergence of applications, so you’ll have high-bandwidth applications, so HD video for example, and that will starve out applications on other devices unless you have a way of virtualizing the RF space and dynamically allocating bandwidth to different devices based on their application need and be able to do that in a scalable fashion as you go beyond five, 10, 20, 30 devices.

And there’s probably the thought that well, I’ll just add more access points without realizing the additional cost that you incur: a) by adding more access points; and b) from the day-to-day management overhead of trying to deal with all of the interference that’s going to happen by adding all of these additional access points.

You know when Steve Jobs had the problem with the iPhone 4 introduction, his problem was all of these people had their Mi-Fis in the audience. Each one was essentially an access point. There were 540 or so access points and nobody could get access to the network.

So what are you seeing in terms of 802.11n adoption? Is that really boosting sales?

We were the first to bring out an 802.11n access point, actually prior to the ratification of the 802.11n spec by the IEEE. And very quickly the majority of our access point sales in terms of number of units became 802.11n. Today I think the number is somewhere north of 95 percent. Probably in a few more quarters people will stop buying 802.11bg. But yes, it’s huge though it’s still in the early days.

Could you lay out for our readers what’s on Meru’s road map?

We have a vertical focus, so we will continue to drive our business in these vertical markets. The single-channel architecture, the virtual cell capability, we will be driving that in an evangelistic mode. We’ve crossed over 4,000 enterprise customers and we’re now in 55 countries around the world, so we’ve crossed into mainstream architecture. So our focus is on methodically and systematically scaling the company in terms of sales resources, channel resources, product roadmap, hardware, software and the like.

I had asked this question of the CEO over at Aruba, but is it difficult to be a stand-alone wireless company when you’re competing against big companies like Cisco that have wireless integrated into a much larger overall architecture?

You know, it’s not difficult if you are offering a compelling value proposition to the customers. And our best prospects are customers who already have Cisco wireless and Cisco wired. So clearly we’re able to go in and solve problems for them or enable applications that either drive revenue for them or drive productivity improvements or drive down costs. Cisco wouldn’t call it this, but I would say we have a symbiotic relationship with Cisco. We help them deliver a better user experience with their applications like CallManager and their wireless voice over IP handsets with Meru wireless than with Cisco’s own wireless.

One of the things that we’ve heard is that as HP gets more aggressive in the networking space, that they need to add wireless capabilities. Have they called you about acquiring you yet?

HP had their own wireless and then they acquired Colubris and then they acquired 3Com, which I’ve heard was data center related. They wanted to grow their market share on the wired side, which is actually a big part of their strategy (How do they grow their ProCurve wired business?). We compete against HP in HP customer locations and I expect that to continue.

Is there anything that we need to know in terms of where you are going with security and management of wireless?

Yes. On the security side, the wireless LAN industry in the early days had a reputation of not being as focused on security as need be. Things like wireless encryption protocol came out and then WPA and WPA 2. And the industry as a whole has made great strides in improving the security of wireless networks. And each company, ourselves included, has invested heavily in delivering not only to the industry standards work but beyond that, such as wireless intrusion protection, PCI for retail and so on. We will continue to bring out features and functions and also products for security.

On the management side, because of the nature of our architecture and the fact that we coordinate in real time between the access points and dynamically allocate megahertz to every device, we capture everything that goes on over the air. So from a management perspective, we have a fairly sophisticated application called E(z)RF and we partner with SolarWinds.

In addition to that, we have introduced an innovation called service assurance manager designed to deliver predictive analytics to enterprise IT. So you can predict ahead of time before you roll out an application, or you can do this at night if you like, where you may have bottlenecks or trouble spots on your wireless network. Our ability to do this is unique. And clearly that’s an area where we will continue to invest.

Read more about wireless & mobile in Network World’s Wireless & Mobile section.