AI Ops: It’s All about Data

All Global 2000 firms are awash in infrastructure data. Every device that an application flows through, be it network equipment, firewalls, servers, load balancers, cloud providers, service providers, storage, etc., provide data to communicate state metrics, events, logs, et al. That is, all the devices that an application depends upon to deliver the intended user experience have some data that provides insight into how it’s operating. The problem is using this data to understand user experience on an end-to-end basis is extremely difficult today as vendors don’t share this data with each other and even when the data is gathered into a central data lake the domain expertise needed is typically broad and distributed. For all the hype in AI for IT Ops, there are very few realistic use cases, thanks to the inability to use infrastructure data across an application’s dependency map. A focus on infrastructure data within the industry can streamline operations with AI. At ONUG Spring in Dallas, we start that process with a group of vendors and IT executives.

In an ideal world, everyone would speak the same language and all infrastructure devices would format their management and monitoring data in a standard language. This standard formatted data could be ingested into a data lake, data store, data warehouse, Amazon’s S3 service, etc. Then AI algorithms could aggregate, correlate, analyze etc., the data and feed a wide range of operational dashboards that provide the status of applications and their dependency map, predicting user experience, capacity planning, anomalistic behavior detection, etc.  While we are slowly moving towards speaking the language via initiatives such as Open Config and NetConf across switches/routers, access points, firewall, etc., this is a slow process. And even when you have all networking equipment speaking the same language you find most problems require deep cross functional domain expertise to understand requiring permission to share that data between vendors.

So at ONUG Spring, a group of vendors and IT executives will start to move the industry’s journey toward AI Ops by presenting a framework for a common infrastructure data store. This data may be stored in traditional silos, but with adding access permission for a firewall, switch/router, WiFi, cloud provider, SD-WAN, service management vendors, etc to each other’s infrastructure data. A common infrastructure data store with selected access permission will enable engineers with domain expertise to correlate this data into useful operational dashboards.

At ONUG Spring, IT business and industry leaders will start an exploratory discussion regarding operational use cases that can be solved with AI if infrastructure vendors work together. For example, if a SD-WAN vendor shares its view of the Wide Area Network connection with other infrastructure vendors, then AI could be leveraged to analyze and predict the end-to-end user experience. Or if a security vendor leveraging AI can share the user risk score with the access infrastructure vendor, then operational teams can monitor or quarantine users based on location and high risk. The goal of the meeting at ONUG Spring in Dallas on May 7th and 8th at the Cityplace Conference Center is to explore use cases/problems IT business leaders need their infrastructure vendors to solve or optimize.

We invited the following companies to participate in the first industry discussion on AI Ops at ONUG Spring. This discussion is scheduled for May 8th at 9:05 am.  

  • Intuit
  • Cigna
  • Citigroup
  • Pfizer
  • State Street Bank
  • Gap Inc.
  • FedEx
  • GE
  • Fidelity
  • Bank of America
  • Velocloud
  • ServiceNow
  • Juniper/Mist
  • Palo Alto Networks
  • Microsoft
  • Amazon

Nick Lippis, Co-founder ONUG, and Bob Friday, CTO Mist, will moderate the discussion.

Join the discussion and add your voice as we plot the future of AI operations for the Global 2000 at ONUG Spring in Dallas on May 7th and 8th.


Author's Bio

Guest Author