A startup coming out of stealth mode sees untapped opportunity in an already crowded application performance monitoring (APM) and network monitoring market to address what is sees are critical gaps in visibility created by the advent of hyper-converged, virtualized data centre.
Chia-Chee Kuan, president and CEO of Santa Clara, Calif.-based Uila, said the virtualized data centre has given enterprises new ways of doing things but it also presents challenges for IT teams tasked to manage infrastructure and keep it stable as they do not have application visibility: “We think that’s a big gap.”
There are in fact two gaps that Uila is looking to address, said Kuan. The first is between the application level and the virtual infrastructure, the latter populated with VMware, KVM and Microsoft Hyper-V. The second gap is between that virtual infrastructure and the actual physical infrastructure made up of compute, networking and storage.
Uila doesn’t see itself as APM tool, said Kuan, as it is more focused on assisting people who manage infrastructure. But in a hyper-converged environment, he said it’s critical to be able to look at all three components of the physical infrastructure. Some tools look at networking but not compute and storage, for example. “We don’t see a tool like ours that covers everything.”
Uila’s taps into network packets from virtual switches to understand which virtual machine is serving which application, as well as other details such as application transaction times, said Kuan. “We analyze every packet that goes in and out of each virtual machine.”
The metadata that is collected gets carried over into an information controller and then to Uila’s cloud-based analytics, although customers do have the option of an on-premises deployment model, said Kuan. The infrastructure team is able to see everything together, from the virtual layer to the physical layer. “The visibility from top to bottom is complete.”
In the past, it would be necessary to get a variety of otherwise siloed team members together when issues occurred and applications were slowing down, said Kuan, but now it’s possible to have all information from an application transaction to quickly pinpoint the cause of the problem as well as predict what may happen in the future to support business-level decisions to optimize infrastructure.
Stephen Hendrick, principal analyst for application development and deployment research at Enterprise Strategy Group, said Uila is not the only vendor thinking this way, but it’s one of the first to have a comprehensive offering in what is a crowded and relatively immature market. “In this space, what differentiates the new breed of players from what’s come before is how much continuous data collection is done.”
While many vendors have taken a log-based approach, Uila is very focused on data packets, noted Hendricks, and although the company doesn’t see itself as an APM vendor, he sees its offering falling into that category, adding that vendors who do APM across the entire stack are able to understand a lot about the network too.
Hendricks said agent-based APM is hard to implement and is usually done selectively for mission-critical applications. Uila’s agentless approach can look at the performance of all applications. “They do packet-level data collection in real-time and do it in a very comprehensive way,” he said. “Looking at the virtual switches gives them a lot of detail. That’s the secret sauce.”
Having holistic, continuous data collection allows an organization to be more data driven, but it’s not easy to do, said Hendricks. And it’s also very early days for all of the various tools designed to monitor enterprise infrastructure, whether it’s application performance or network behaviour. Down the road, he could see data collection and data analysis as two different disciplines that vendors could specialize in, but for now most are going to purport to offer one-stop shopping.