Mike Matchett of Small World Big Data and Scott Baker of IBM discuss the latest trends in IBM storage, including cyber resilience, AI operations, and sustainability. Scott highlighted IBM's focus on providing the best cyber resilient storage in the market. One of their innovations allows for real-time data corruption detection in virtual machine environments, using machine learning to monitor how data is written and the interaction between the host and storage array. This system can spot potential cyber threats by tracking anomalous behavioral changes.
Scott also emphasized the importance of operationalizing AI in storage. Until organizations can operationalize AI, monetizing it remains a challenge. The future of storage will demand innovative computational storage requirements and an ecosystem that provides data storage, as well as services that extract or add value to the data. This is crucial for processing in AI workflows and ensuring proper governance.
Mike pointed out the growing intelligence of storage systems, which now use AI to detect irregularities in data storage patterns. He questioned Scott about the future of workload management. Scott explained that AI requires performance intensity and capacity sensitivity. Successful AI storage solutions should offer a wide ecosystem of interoperability, support various data and application types, and provide both dedicated appliances and software-defined options. The conversation also touched on the importance of a holistic ecosystem that includes consultation, compute, and networking layers. Scott foresees a need for elasticity in infrastructure, where GPU farms and capacity can be dynamically adjusted based on immediate needs.