Aruba succeeded where other Wi-Fi companies failed: A talk with the founder about the acquisition by HP, what comes next with Wi-Fi
- 15 October, 2015 19:42
Wireless LAN stalwart Aruba was acquired by HP last March for $3 billion, so Network World Editor in Chief John Dix visited Aruba co-founder Keerti Melkote to see how the integration is going and for his keen insights on the evolution of Wi-Fi. Melkote has seen it all, growing Aruba from a startup in 2002 to the largest independent Wi-Fi company with 1,800 employees. After Aruba was pulled into HP he was named CTO of the combined network business, which employs roughly 5,000. In this far ranging interview Melkote talks about product integration and rationalization, the promise of location services and IoT, the competition, the arrival of gigabit Wi-Fi and what comes next.
Why sell to HP?
Aruba was doing really well as a company. We gained market share through every technology transition -- from 802.11a to “b” to “g” and now “ac” -- and today we’re sitting at roughly 15% global share and have a lot more than that in segments like higher education and the federal market. But we were at a point where we could win more if we had an audience at the CIO level, and increasingly we were getting exposed to global projects that required us to have a large partner in tow to give us the people onsite to execute on a worldwide basis.
So we began looking for what internally we called a big brother to help us scale to that next level. We talked to the usual suspects in terms of professional services, consulting companies, etc., but then HP approached us and said they were interested in partnering with us to go after the campus market, which is changing from wired to wireless.
HP has a good history on the wired side, so we felt this was an opportune moment to bring the sides together, but go to market with a mobile-first story. After all, as customers re-architect their infrastructure they’re not going with four cable drops to every desk, they’re looking at where the traffic is, which is all on the wireless networks these days. HP agreed with that and basically said, “Why don’t you guys come in and not only grow Aruba, but take all of networking within HP and make it a part of the whole ecosystem.”
So HP Networking and Aruba have come together in one organization and Dominic Orr [formerly CEO of Aruba] is the leader for that and I am Chief Technology Officer. We are focusing on integrating the Aruba products with the HP network products to create a mobile-first campus architecture.
Does the Aruba name go away and does everyone move to an HP campus?
No, and there is some exciting news there. The go-forward branding for networking products in the campus is going to be Aruba, including the wire line products. Over time you will start to see a shift in this mobile-first architecture with Aruba switching also coming to market.
And we talked about real estate. Do we all move to the HP facilities in Palo Alto or do we stay here? The good news is we’re going to stay here for at least 12 months, and after that we are building a facility for Aruba, so we are going to move into a new building right down the street.
Will that include the HP Networking operations in the area?
No, we have a global development model, so we have development sites here in Sunnyvale, Palo Alto and Roseville. And we have sites in India, China, Canada and in Costa Rica. There won’t be any changes to any of the development sites. As the business grows we’re going to have to grow most of those sites.
HP has bought other wireless players along the way, including Colubris and 3Com, so how does it all fit together?
Colubris was a pretty focused wireless acquisition back in 2008 and those products have done well for HP, but that customer base is ready for upgrades to 11ac and as they upgrade they will migrate to Aruba. The former product line will be end-of-lifed over time, but we’re not going to end support for it. There is a small team supporting it and will continue to do so until customers are ready to migrate.
3Com was a much broader acquisition, involving data center campus products, routing, etc. Most of the RD for 3Com is in China with H3C [the joint venture 3Com formed with Huawei Technologies before 3Com was acquired by HP in 2010]. There is a two-prong go-to-market approach for those products. There is a China go-to-market, which has done really well. In fact, they are number one, even ahead of Cisco, from an overall network market share perspective in China. For the rest of the world we were using the products to go after the enterprise.
As you probably heard recently, we are going to sell 51% of our share in H3C to a Chinese owned entity because there needs to be Chinese ownership for them to further grow share. H3C will be an independent entity on the Chinese stock market and will sell networking gear in China and HP servers and storage as well.
So that becomes our way to attack the China market while we will continue to sell the other network products to the rest of the world. Those products are doing very well, especially in the data center. They run some of the largest data centers in the world, names that are less familiar here in the U.S., but very large data centers for the likes of Alibaba, Tencent and other companies that are basically the Amazons and Facebooks of China.
3Com has a wireless portfolio called Unified Wireless. That product line will also be end-of-lifed but still supported, and as we migrate to next-generation architectures we will position Aruba for those buyers. The definitive statement we’ve made is Aruba will be the wireless LAN and mobility portfolio in general and Hewlett-Packard’s network products will be the go-forward switching products.
Two products that are really helping to integrate our product lines are: ClearPass, which is our unified policy management platform, which is going to be the first point where access management is integrated between wired and wireless; and AirWave, which is the network management product which will become the single console for the customer to manage the entire campus network. For the data center we will have a different strategy because data center management is about integrating with servers and storage and everything else, but for the campus the AirWave product will be the management product.
3Com has a product called IMC Intelligent Management Console that will continue if customers need deep wired management, but if you need to manage a mobile-first campus, AirWave will do the complete job for you.
Given your longevity and perspective in the wireless LAN business, are we where you thought we would be in terms of Wi-Fi usage when you first started on this path 13 years ago?
It’s taken longer than I thought it would, but it has certainly far surpassed my expectations. Back in 2002 there was no iPhone or iPad. Wireless was for mobile users on laptops and we believed it would become the primary means of connecting to the network and you would no longer need to cable them in. That was the basic bet we made when we started Aruba. My hope was we would get there in five to seven years and it took 15, but things always take a little bit longer than you think.
The seminal moment in our business was the introduction of the iPad. Even though the iPhone was around most people were still connecting to the cellular network and not Wi-Fi because of the convenience. Laptop-centric networking was still prominent, but when the iPad arrived there was no way to connect it to the wire and there were all sorts of challenges. How do you provide pervasive wireless connectivity, because the executives that brought them in were taking them along wherever they went. Security was a big challenge because they were all personal devices.
We had developed and perfected answers for those questions over the years so it was all sort of right there for us. And the last five years has seen dramatic changes in terms of all-wireless offices, open office space architectures, etc. Microsoft Lync was also a big inflection point as well.
Why is that?
Whenever I talk to customers about pulling the cable out they always point to the phone and say, “I still need to pull a cable for that, which means I need power over Ethernet, I need an Ethernet switch in the closet, I need a PBX.” But when Lync was introduced in 2013 you could get your unified communications on your smart phone. Today, if you were to ask what is the most important device on the network, I’d say it’s the smart phone because it’s converging the computing and messaging and everything else on one device. Now you can provide a rich experience on a mobile device and do it anywhere, anytime.
Where do we stand on location-based services?
We’ve been talking about location services for a very long time. What happened was Wi-Fi based location alone wasn’t actually solving the problem. It was giving you a sense of where people were in a facility, but getting the technology to allow you to engage with somebody in physical space was not working, mostly because the operating systems on those mobile devices weren’t supporting Wi-Fi for location, just connectivity.
We have now integrated Bluetooth Low Energy (BLE) into our portfolio so you have two ways of connecting with the user; the Wi-Fi side gives you presence and Bluetooth Low Energy gives you the ability to engage on the user side so you can send notifications about where they are. That technology lets us provide tools for marketers, for retailers to send coupons, invite people into a store, and so on.
So it is finally picking up some?
It is. Actually Asia is doing well. There is a lot of construction in Asia and this is one of the demands. But the U.S. is picking up. We just implemented a large network at the Levi’s Stadium right down the street here [which recently replaced Candlestick Park as home of the San Franciso 49ers].
One of the things the CEO imagined was that, as you drive from home to the game, their app would guide your experience. So they’ll take you to the right parking lot, then provide you directions to your seat, and once you are in the seat enjoying the game they wanted to provide amenities -- so food and beverage ordering and the ability to watch instant replays and the like. All these things are available for a fee of course. In the first season of operation this app generated $2 million of additional sales for Levi’s Stadium.
That was a big win for us, not just for demonstrating high density Wi-Fi where we have seen regularly 3-4 gig of traffic going to the Internet, but also showing the revenue generating potential of location-based technology.
Speaking of networking a lot of things, what do you make of the Internet of Things movement?
Eventually where it all goes is integrating the Internet of Things. Every day I interact with customers there are new use cases coming up around the intersection of location-based technology and the Internet of Things. And that’s squarely in the purview of what we are doing. It’s not today. Today is still about this all-wireless workplace, but in the next five years I think you’ll see a lot more of this. There is a lot of innovation still to come.
There’s a hodgepodge of stuff used to connect sensors today, but you see Wi-Fi playing a prominent role?
Wi-Fi will definitely be an integral component, but Bluetooth Low Energy will also be important because some sensors will be battery operated. There may be a role for the evolution of ZigBee as well. That’s super low energy. ZigBee is not yet in the mainstream enterprise but I can see some successor of that happening. But sensors will look to wireless for connectivity because they need to go anywhere. You can’t have cable follow them. So the wireless fabric is becoming super-critical for that.
Switching gears a bit, how is competition changing?
We look at three key market segments: large and medium enterprises; small/medium businesses, which have completely different characteristics; and service providers. Aruba has done really well in the large and medium enterprise segment. We have done reasonably well in the small/medium segment, but there is more competition there. Ruckus has done well there. And service provider is the emerging battleground.
As a standalone company Aruba couldn’t afford to invest, frankly, in all three segments. We were focused on the large and medium enterprise and we built a good franchise. Clearly Cisco is the primary competitor there, but now as part of HP we have another go-to-market capability and investment to take on all three segments in a meaningful way, so that’s another big reason why we came together.
We just recently announced a partnership with Ericsson to go after the service provider Wi-Fi segment, and that will help us gain share. And HP has been a strong player in the small/medium business so we’re going to take Aruba down-market. We’re going to play in all three segments. I feel if we just keep executing, market share gains are possible.
Ruckus talks about optimizing the airwaves as being their key differentiator. How do you differentiate Aruba?
The four key things I talk about are the emergence of the all-wireless workplace, inflight communications and voice, the need for deep security within your own device, and the need for location based services and training towards IoT.
We talked about the all-wireless workplace and location services. Regarding voice traffic, we have invested quite a bit of energy ensuring optimal utilization. Ruckus focused on the antenna technology, while we are focused on the software that goes on top of the antenna. The analogy I’ll give you is, as you walk away from an access point I can boost my antenna power to give you a better signal, and that problem is a good problem to solve if you’re in a home because you only have one access point. But in the enterprise there is a collection of access points and the problem isn’t about holding onto a client for as long as possible, but to move the client to the best access point. So the trick is to enable the client to roam from one access point to another in a very efficient way. We call this technology ClientMatch. That is the core differentiator for us over the air, and we’ve specifically optimized it for voice by working with the Microsoft team to enable Lync and Skype for Business.
Security is a place we cannot be touched. We’ve had deep security expertise for a very long time. The DoD, three of the armed forces, most of the federal market actually, uses Aruba. I can’t get into all the details, but we have significant penetration because of our security depth. For enterprises that is a big deal. They really want to make sure the security side is well covered.
What’s the hot button in wireless security today?
We know how to encrypt. We know how to authenticate. Basically it is the threat of an unmanaged device coming into the network. We’re looking at solving that problem as a mobile security problem and we solved one part of it with access management, but we have this Adaptive Trust architecture which integrates with mobile device management tools -- VMware’s AirWatch, MobileIron, Microsoft’s Intune. We partner with those companies and the likes of Palo Alto Networks, and HP now brings its security and management platform ArcSight to the table. The idea is to secure the mobile edge so no matter where you are you have a secure connection back to the enterprise.
Let’s shift to the adoption of Gigabit Wi-Fi, or 802.11ac. How is that transition going?
The campus access network from your desktop to the closet has stagnated for a long time. That’s because there was really nothing driving the need for more than a gigabit’s worth of bandwidth to the desktop. Now with Gigabit Wi-Fi technologies the over the air rates are greater than if you were to connect to the wired LAN. So if you deploy Gigabit Wi-Fi and have signals going at 2G, let’s say, the wired line becomes a bottleneck. There is a technology called Smart Rate that HP Networking introduced for its switches which allows you to raise the data rates to 2.5Gbps and even 5Gbps. At that point your access points don’t have to contend with the bottleneck and can pick up the bits over the air and put them on the wire without dropping them.
So you need will need wired ports faster than a gigabit as you transition to this mobile workplace, but you won’t need as many ports as before. That is a transition, I think, that will happen over the next 2-3 years.
Did many people buy into Wave 1 of Gigabit Wi-Fi or did they hold off?
We’ve had tremendous success with Wave 1. The need for bandwidth is truly insatiable. And there is a ton of demand still yet to be put on the network. Video is a significant driver of bandwidth and most companies are throttling it video. So the more you open the pipe, the more capacity I think people will consume. Wave 1 has done very well. I think Wave 2 will continue to do well and then there’s .11ax which will take capacity even higher.
So people bought into Wave 1 even though Wave 2 requires them to replace hardware?
I tell customers, if you’re going to wait for the next best thing you’re going to wait forever, because there’s always going to be the next best thing on the horizon. So it’s really a question of where you are in your lifecycle for an investment. If the customer is at a point where they’ve had five years of investment and they’re hurting, it’s a good time. Wave 1 can actually solve a lot of problems. There’s no need to wait another 18 months for Wave 2 technology. You know you’re going to refresh that too in five years and there will be new technology at that point in time.
Will anybody buy anything but Wave 2 at this point?
It depends. Wave 1 technology you can buy today at multiple price points in the industry. Wave 2 is still at the very top end of the range. So if you’re looking for, let’s say, lighting up a retail store and you don’t need all the capacity of Wave 2, then Wave 1 will do just fine. That’s typical of most technologies, to start at the top and eventually work its way down. We’re right in the beginning of the Wave 2 transition.
How about in carpeted office space? Would you just drop Wave 2 into key points to satisfy demand?
Wi-Fi has always basically been single user. Only one user could speak on a wireless LAN at a time. With Wave 2 you can have multiple conversations at the same time; each access point can serve four streams. So that boosts capacity in a significant way and can also improve spectrum efficiency. For that reason alone, I think Wave 2 should be used pretty much anywhere you go. You could start with a high density zone and then work your way up. That’s typically how people do it, but I would encourage most customers to take advantage of this technology.
In the industry we’ve always used speed as a measure of the next generation of technology. Never have we given attention to efficiency. This is the first time where we’re saying efficiency gains are pretty significant.
And Wave 2 ultimately will be able to support up to eight streams, right?
Yes, the technology allows you to do eight streams, although it is not possible to pack eight antennas into the form factor at this point. But it will come.
You mentioned 11ax. What is that about?
This is still a nascent technology. It’s not yet standardized, but it is being rolled into the IEEE.
I think the targets are up to 10 gig. Let’s see how far they get. At that point, the Gigabit Ethernet backhaul will become an even more interesting problem. You’ll need 10 gig of backhaul from the access point.
In terms of the coming year, what should people look for?
They should expect a streamlined roadmap with unified management for wired and wireless, and unified security for wired and wireless in the campus. And they should expect changes in wiring closet switches to support Wave 2.
The other piece cooking in the labs is the next-generation controller technology. We invented the controller back in 2002 and that has gone through multiple generations of upgrades. The first controller had something like a 2Gig back plane that could support 1,000 users, and now we have a 40G controller that supports 32,000 users. So how do you get from there to 500G? That will require us to rethink architecture because these campuses are getting there.
We used to talk about tens of thousands of devices on a campus. Today campuses have hundreds of thousands of devices. How do you support them in a single architecture? Right now you add more controllers, but that creates a management problem. We are working on a unified solution for very large campuses and taking it to the next level for service providers as well.