Cisco this week announced its intent to acquire Perspica, a machine learning vendor -- the second of this sort of acquisition this year (not to mention its 200th overall).
Earlier this year Cisco dropped a hefty $3.7 billion to purchase AppDynamics, paying a significant premium over the application performance vendor's pending IPO price. AppD, as it's better known as, gave Cisco a view of user experience through the lens of applications. Prior to the acquisition, Cisco had to infer how apps were performing through data generated by its security, networking, and data center products. AppD gave Cisco badly needed application intelligence, arguably enabling it to deliver the first true end-to-end monitoring and analytics platform.
Modeled for Scale and Speed
To provide enterprises a view of how applications are performing, AppD analyzes and applies machine learning to the data it collects from users, applications, databases, servers, and, soon, networks and Internet of Things endpoints. It aggregates collected data and works on data at rest. That's not to say that AppD doesn't provide a real-time view, because it certainly does. However, it does need to combine and sort the data.
Conversely, Perspica uses machine learning to analyze real-time streaming data, or data in motion. This could be the telemetry generated from many of Cisco's newer network products, its ACI controller, or Tetration analytics system, as well as any competitive products that stream data. Perspica can process enormous volumes of data on the fly. Together with AppD's product, the Perspica machine learning tools will enable Cisco to measure business transactions at greater scale and speed than it is able to do now.
The n-tier concept has been around as long as applications have been, but the "n" used to be something countable on one hand. Now "n" possibly means dozens of tiers that are constantly changing as virtualization and containers enable processes to spin up instantly. The application management platform created from AppD and Perspica technologies will model business transactions at such scale and speed that it'll be able to instantly inform business leaders, application developers, and IT operations or network engineers of any issue that might impact user experience, as well as allow them to quickly isolate the source of the problem.
Promising Opportunities
Cisco raised many eyebrows by paying what many perceived as a steep premium for AppD. To me, purchase price is somewhat irrelevant. More important is the acquisition's success. For example, if Cisco had paid 10 times what it did for Selsius Systems to move into VoIP back in the late '90s, no one would have batted an eye. That purchase opened up a market adjacency, bringing Cisco new buyers and channel partners and pulling through billions of dollars of networking products. On the other hand, if Nortel had paid one tenth of what it did for Alteon WebSystems back in 2000, we would still look back on it as a failed acquisition.
AppD presents some interesting opportunities for Cisco, which has been trying to move "up the stack" for years and become more relevant to CIOs and lines of business (LOB). It's been successful in making that happen in a few areas, like smart cities, but in general Cisco doesn't have much to create a talk track with business leaders. Cisco salespeople have often told me that access isn't the problem; they can get meetings with business executives, but they don't have much to say to them as no one outside networking cares about things like 60-Watt Power over Ethernet or 2.5-Gig networking, and its arguable that network operations really don't matter to them, either.
AppD gives company something to create a talk track, as the solution has direct impact on the business. I had the chance to talk to some AppD customers, such as United Airlines and Carhartt, this week at an AppD Summit Cisco held in New York. At those companies, application developers, IT operations, and LOB managers use the AppD platform. AppD has become core to making sure these companies can quickly identify and remediate any problem that might impact worker or customer experience.
Over time, if Cisco is able to build the right capabilities into its networking products, it could definitively say that AppD on a Cisco network will make apps run better and faster than they do on a competitive network. That kind of value proposition could create some pull-through for a networking business that has been in decline for years. Businesses aren't in a rush to upgrade their networks for the sake of an upgrade, but they might if they know the core applications that power their businesses will run better and more securely.
From a sales perspective, I think Cisco should be leading with AppD as much as possible, since everyone cares about application performance. Despite the high price point, AppD was a good acquisition for Cisco. Obviously Cisco agrees, since it's adding even more value through Perspica.
Related post:
Follow Zeus Kerravala on Twitter and Google+!
@zkerravala
Zeus Kerravala on Google+