News from HPE’s Nonstop Division

It is now officially impossible to ignore the AI storm that is shaking the IT industry. Many of us have been through previous major industry shifts, such as the rise of the world wide web in the 90’s, the mobile computing, the cloud computing but none have challenged the way we think, design, build, deliver, and maintain platforms and software as AI. As much as I want this newsletter to cover various topics, I suspect AI and its impact on Nonstop is going to be our main discussion for the time being.

As it happens, major shifts like this are usually very bumpy at first. Indeed, for AI, after the democratization of machine learning and the introduction of generative AI, the AI ecosystem is not only moving fast but also changing directions. One recent change has a lot to do with the introduction of the MCP standard often described as the USB-C for AI or the new REST API. It is indeed providing a standard for AI agents to interact with existing software, APIs and tools and has led to the introduction of the term “Agentic AI”. This is a change in direction because up to now the trend has been to bring AI into existing software, AI embedded inside your tools, the copilot model, massively adopted, such as Microsoft 365 Copilot. While providing great assistance within each of those tools, this approach lacks the end-to-end workflow view.

Agentic AI and the MCP model flip the paradigm. Instead of embedding AI in each and every tool, you instead let existing tools communicate via the MCP interface to an overarching AI assistant. Such assistant is able to communicate with multiple MCP servers, can retrieve real-time data from multiple sources, as well as operate multiple tools and therefore coordinate a full workflow.  The example of “Claude code” is a perfect example of the new paradigm being very successful in a DevOps context where the AI assistant can execute all the steps of a typical DevOps tools chain such as editing, compiling, testing and deploying. In a context where AI is challenged with multiple projects failing or not bringing the expected return on investment, there is pressure to deliver more value. Agentic AI may prove to close that gap.

Beyond this context, the Nonstop community has also been identifying the areas where AI would bring most value. This edition has perfect examples showing how Nonstop is going to fit into the AI ecosystem going forward. First an article from Infrasoft introducing a very solid MCP server implementation which, as we pointed out, is an essential enabler for agentic AI. Then an announcement from Xypro about their AI assistant “Lionel” leveraging knowledge of Nonstop institutional knowledge to more quickly solve complex security and compliance issues. And third, the TIC software Navigator, addressing a well-known pain point for preserving the business value of code that is becoming opaque yet still extremely critical to customers’ business.

So yes, AI is good for Nonstop. MCP is going to make Nonstop manageable like any other platform, abstracting proprietary interfaces, yet keeping the Nonstop fundamentals’ benefits in the background. AI is going to help with code maintenance helping preserve many years of investments. AI is going to strengthen and streamline operational, security and compliance workflows. Good news for Nonstop.

Enjoy this new edition!