Epsilon Talks Using AIOps to Streamline Asset Management

Modernizing the legacy infrastructure at Epsilon, a provider of email marketing services and other resources to brands, accelerated quickly when the company embraced cloud-based AIOps. CIO Robert Walden says that four years ago, Epsilon ran on-prem datacenter hosting but saw the massive growth taking place in public cloud. He says the company crafted some of its own IT resources for a time but faced an escalating need for multiple solutions to resolve multiple demands around asset management and monitoring. “We had a program we put together called Orion,” Walden says, to help the company get in front of things.

Epsilon looked at several different products, he says, but the products available at the time from vendors tended to come from repurposed, older resources. “Especially the bigger players, they cobbled together different solutions, bolted them into each other and called it something new,” Walden says. The company wanted a platform that took modern demands into consideration, he says, and fulfilled Epsilon’s pursuit of robust asset management. “We needed to capture a lot more detail, a lot more in-depth information and be able to attach metadata,” he says.

Many asset management solutions on the market at the time were built for traditional, internal IT shops, Walden says, rather than service providers, further complicating matters. Epsilon needed a resource that could meet the challenge that came from using multiple client solutions and product platforms that are domains within themselves. Walden says separating and segregating information into their respective domains is useful, which is what he says OpsRamp, a provider of AIOps and cloud monitoring, brought to the table.

Visibility, understanding who was doing what and where something happens in a system, was one of the key starting places for managing an environment such as Epislon’s, says Radhakrishna Kukunuri, solutions architect for OpsRamp. “Providing perspective to the right stakeholders on a massive scale was a complex and daunting task for [Epsilon],” he says. OpsRamp built multitenancy into its design, Kukunuri says, which fit Epsilon’s needs as a distributed marketing solutions provider with dedicated servers for each customer.

Epsilon wanted to put together a view for each respective business unit, he says, that let each customer see only their respective assets. “They had a challenge in how they can share a very siloed view for each business unit,” Kukunuri says.

The pandemic obviously changed up certain demands, such as driving organizations to optimize at scale with cloud-based solutions while working with limited teams, he says. “That amounts to more automation required in the IT operations segment.” That also drove the need for data to be processed in smarter ways when presented to engineers.

Walden says as Epsilon grew quickly, it worked with the OpsRamp to incorporate its needs and requirements into the roadmap of the resources being developed. “We needed to have this solution in place three years before we started,” he says.

Image: Robert Walden, Epsilon

Epsilon’s transformation has seen incremental improvements, he says, not just sweeping alterations. Those changes in process included going from manual to semi-manual to automated to fully automated to optimized. For example, Epsilon altered how it handles notifications to better identify issues, the platform the issue arose on, and who needs to be involved to resolve the matter, Walden says, which might otherwise take an extensive time. He says OpsRamp eliminated that process and sends the right messages at the right time and place.

Additional benefits Walden says Epsilon found with OpsRamp include opening up more time for other value-added activities beyond AIOps. “We had a major cultural change,” he says. “We had to make the move from traditional on-premise hosting to hybrid hosting in the public cloud.” Those shifts included Epsilon’s IT team looking at things as code and not just as traditional infrastructure. This evolution in understanding meant making use of new tools to facilitate and execute the cultural change. “It really became part of our overall transformation to this modern world of operations,” Walden says. “It became a fundamental part of our strategy.”

For more content on AIOps, follow up with these stories:

How AI and Machine Learning are Evolving DevOps

AIOps Vendors Should Prepare for a Litmus Test in 2020

How AIOps Can Improve Data Center Management

Moogsoft and Cisco’s AppDynamics on How AIOps is Growing