Earlier this year in February, Brocade announced its acquisition of Riverbed’s SteelApp line of products just one month before its announcement that it would acquire Connectem, a company in the LTE virtual evolved packet core (vEPC) market. Both moves are the latest in a recent series of acquisitions underway to strengthen Brocade’s offerings in software-defined networking (SDN) and virtualized network functions. SDxCentral’s Roy Chua sat down with Brocade Vice President of Service Provider Strategy Andrew Coward to talk about how Brocade’s SDN & NFV acquisitions support its larger vision for the service provider market.
SDxCentral: Brocade has been quite busy since last fall, when you began the acquisition of Vistapointe. Since then, Brocade bought Riverbed Technology’s SteelApp line, and Connectem. How would you describe Brocade’s overall acquisition strategy?
Coward: Our objective is not to solve every problem in the entire ecosystem, but to bring unique technology to market that solves real business problems in a way that isn’t happening anywhere else. There is no point in having a service provider strategy that competes head-on with traditional vendors with traditional products. As we look at acquisition opportunities, we look at what problems haven’t been solved yet that should or could be solved, but aren’t yet being addressed by the industry in a meaningful way.
How has Brocade’s presence in the service provider space evolved? How do you use acquisitions to strengthen your position?
Coward: Our strengths in the service provider space are rooted in our acquisition of Foundry in 2008, particularly their routing products that are deployed in many large service providers and exchange providers today.
The Vyatta acquisition in 2012 gave us a new angle to compete and put us on a road toward software networking even before NFV came on the scene. The Vyatta product was ideally suited to address many of the business opportunities that eventually emerged from ETSI use cases like virtual routing, virtual CPE, virtual firewall, and so on. Vyatta allowed us to start focusing on cloud service providers, and we began to build our larger service provider organization from there.
Having built a global service provider team, we’ve expanded the product portfolio and business value we deliver. The common, compelling value proposition in all our acquisitions is the performance attributes that we can extract from these products. The explosive growth of traditional mobile data and all the new requirements around Internet of Things (IoT) and machine-to-machine (M2M) connectivity is placing extreme demands on mobile networks. Service providers know they need to leverage cloud computing, network virtualization, and software networking technologies but it’s really hard to take legacy operating systems from traditional routing manufactures, squeeze them into a server, and expect to run tens or hundreds of instances across that server infrastructure.
All of our acquisitions are geared toward leveraging x86 technologies to give providers extremely high performance that matches or exceeds traditional proprietary products. We aim to provide our products in a form factor and software size that enables providers to deploy these network functions millions of times and do so really efficiently. These products fit broadly into the New IP industry initiative that is designed to help move service providers from closed, proprietary hardware to an open, software-based eco-system.
How will Brocade’s more recent acquisitions help providers leverage next-gen technologies?
Coward: The acquisitions give us the opportunity to solve very specific problems providers are facing today. Many of our customers have started establishing an SDN element in their networks, which is the critical first step. Then it becomes a question of what services can we pull off the existing infrastructure and existing hardware, and pull them into software elements where they became more efficient, cheaper to deploy, and more flexible because the customer gets control of them. And again, how do we do that to solve problems in a way no one else is doing?
For example, Connectem takes an approach to solving the packet gateway problem in a very different way. Why do we have to end up processing packets in four or five different locations for different functions? That doesn’t make any sense at all. Connectem consolidates packet gateway functions to process each packet just once, which is much more efficient. This is going to be important as the world moves toward IoT because the cost point of processing a packet has to be way, way lower than it is now, especially as we think about the next billion users who have yet to be connected, and the billions of devices after that.
With the Riverbed SteelApp software acquisition, the SteelApp product line had a solid track record of controlling traffic to and from applications using x86 software and enabling fast, reliable, secure application delivery. SteelApp has a wide installed base of product, which we are now leveraging to deliver multiple software networking functions.
Finally, Vistapointe helped solve a next-generation analytics piece that no one seemed to be addressing, which is the problem of virtual probes and virtual packet brokers. One of our Asian service provider customers said it was deploying virtual packet gateways, but had “lost its packets” on the way from hardware to software. Providers need to understand what services and what applications are going across the network.
Rather than looking at traditional voice quality, which isn’t as relevant today because it has been solved, Vistapointe looks at things like real-time application performance. Our providers can start asking questions like, “Do customers get a better experience with VoLTE or with Skype?” or “What’s causing network congestion in that cell tower?” The Vistapointe product also enhanced capabilities of our existing packet broker product because Vistapointe can instruct the packet broker using SDN technologies in real time about which streams to direct to which system. These two things come together in a way that makes total sense.
Did any of your acquisitions open the door to another?
Coward: The sequencing of the acquisitions we’ve done last fall and early this year was just a matter of opportunity. That said, it’s exciting to see how they will be coming together to solve problems for service providers. For example, going back to the load balancing piece from SteelApp, there are a number of interesting problems that load balancing really applies to. I’m thinking largely of resilience and redundancy, which are really important to all the services being delivered.
Resilience and redundancy may sound like very pedestrian types of problems to solve for, but if you think about it, how do you sense that a server has failed or an instance of the virtual router or a VPN gateway has failed or that a firewall is overloaded? How do you make sure traffic is immediately passed to a new version of the network function and make sure that instance comes up to form part of the new load balancing equation?
Those problems are quite hard to solve from a traditional routing perspective, so load balancers have come into their own because they are in turn redundant and effectively provide an infinite number of fallbacks. In tying together some of the technologies we have within the Vyatta product that provide stateful failover, we now have a way of directing traffic so that an IPsec connection never goes down. The architecture behind this is what brings more resiliency to networks.
How does Brocade’s work with OpenDaylight tie into your acquisition or service provider strategy?
Coward: Through our work with OpenDaylight, we’re finding the number one problem service providers have is bringing the old and the new together, or basically automating the old with the new. When SDN and NFV initiatives first started out, service providers were imagining the network in five or ten years’ time. Now, as carrier projects have become more real, to the real issue is how to bring the physical and virtual worlds together.
Because OpenDaylight provides a “universal translator” between the business logic of a service, and the actual configuration required to deploy the service, it plays a vital role in bridging the old with the new.
Our release of the Brocade SDN controller is a Brocade supported, quality-assured edition of the OpenDaylight controller code. By providing a commercial release of OpenDaylight, we’re solving the translation problem and a whole host of others. We’re very keen on making sure the industry keeps the management layer as open as possible. We think there’s a grave danger of the industry being locked in again should those management tools be dictated by traditional vendors with proprietary software. Industry initiatives such as OpenStack and OpenDaylight are key to the health of the industry in keeping vendors open and (from a customer perspective) easily replaceable.
To the extent we can contribute to projects that facilitate the openness of these foundational network technologies, it benefits us, it benefits customers, and in the long term, it will benefit the industry as a whole.
Thank you for your time today.
Coward: Thank you for having me.