{"slug": "storage-news-ticker-28-october", "title": "Storage news ticker - 28 October", "summary": "Adeptia released Automate 5.3, an agentic platform update that expands native Model Context Protocol capabilities so users and AI agents can build and validate data mappings in natural language, cutting onboarding time from days or weeks to minutes. The release also adds Bring Your Own LLM support and Observe platform-health monitoring. In the same 28 October storage news ticker, Chelsio launched its Seventh-Generation AI Interconnect Platform with S7 SmartNICs, T7 Storage Controllers and T7 DPUs supporting native 400GbE and Unified RDMA, and CloudCasa by Catalogic added MCP server support, Dell PowerStore disaster recovery and a preview of Changed Block Tracking VM backups.", "body_md": "# Storage news ticker - 28 October\n\nETL supplier **Adeptia** announced Automate 5.3, its latest agentic platform release. Automate 5.3 expands its native MCP capabilities, enabling users and AI agents to create [data mappings](https://www.youtube.com/watch?v=SQpfxPrwWzM) using natural language and apply context-aware insights. This reduces reliance on deep technical expertise, speeds partner onboarding, and makes building new integrations easier at scale.\n\nThere is a redesigned AI Mapping experience that lets developers, business users, and AI agents build and validate mappings in natural language, cutting onboarding time from days or weeks to minutes. The AI learns from each mapping and becomes increasingly familiar with the organization’s data, terminology, and business rules. 5.3 also includes updates such as Bring Your Own LLM support and Observe now for monitoring platform health.\n\nLearn more at an [upcoming webinar](https://edge.prnewswire.com/c/link/?t=0&l=en&o=4806460-1&h=1940861067&u=https%3A%2F%2Fadeptia.zoom.us%2Fwebinar%2Fregister%2F2917924182448%2FWN_UnCI3rAKSRCWgxFYQmj-Fw&a=join+our+upcoming+webinar), Meet Your New AI Mapping Agent: What's New in Adeptia 5.3.\n\n…\n\n**Chelsio** announced its Seventh-Generation AI Interconnect Platform, introducing a comprehensive portfolio of SmartNICs, Storage Controllers, and Data Processing Units (DPUs) designed to accelerate AI training, inference, storage, and cloud networking. Built on the proven Unified Wire architecture, the platform delivers native 400GbE, Unified RDMA (iWARP and RoCEv2), advanced storage acceleration, and AI infrastructure optimization while preserving software compatibility across previous Chelsio generations.\n\nThe AI Interconnect Product Family platform spans three product categories built on Chelsio's seventh-generation offload architecture: S7 SmartNICs, memory-free single-chip solutions optimized for cloud, virtualization, edge, and AI deployments; T7 Storage Controllers, dedicated storage acceleration adapters offloading NVMe/TCP, NVMe-oF, iSCSI, and RDMA workloads; and T7 DPUs, fully programmable adapters that accelerate AI networking, storage, and security with native support for up to 400GbE.\n\nThere is more information [here](https://www.chelsio.com/wp-content/uploads/resources/pr-CHELSIO_T7_LAUNCH_FINAL_V2.pdf).\n\n…\n\n**CloudCasa** by Catalogic announced a feature update, led by CloudCasa Disaster Recovery (DR) support for an additional enterprise array, new Model Context Protocol (MCP) support that allows agents and AI tools to interact with CloudCasa, and a preview of Changed Block Tracking (CBT) support for virtual machine backups. The updated CloudCasa feature set is available now, with CBT backups in preview. In a tad more detail;\n\n- Storage-Integrated Disaster Recovery. Dell PowerStore now joins HPE Alletra Storage MP B10000, HPE Alletra 9000, and SUSE Storage as supported storage replication providers for CloudCasa DR. With CloudCasa DR, the storage system replicates persistent volume data to the recovery site, while CloudCasa orchestrates DR failover and failback of Kubernetes applications and KubeVirt VMs. Details are available for Dell PowerStore and HPE Alletra.\n- Model Context Protocol (MCP) Support. CloudCasa now provides an MCP server, enabling AI agents to easily access and interact with CloudCasa’s capabilities. Platform teams can now monitor and report on their data protection and DR status using agents and other AI tools compatible with MCP.\n- Changed Block Tracking Preview. CloudCasa has added support for virtual machine backups using CBT, to significantly shorten backup times. CBT identifies and logs which specific disk blocks have been modified since a previous backup.\n\nMore details [here](https://docs.cloudcasa.io/help/relnotes-09-2026.html).\n\n…\n\nEnterprise risk intelligence across capital markets supplier ActiveViam has a collaboration with **Databricks** to deliver validated, real-time risk analytics built directly on the Databricks  Data + AI Platform, helping financial institutions build and deploy AI agents grounded in governed enterprise data and trusted risk calculations.\n\n…\n\nUNSW Sydney is using **DDN** storage in its Research Data Experience (RDE) program. **Arcitecta** will combine its MediaFlux data management software with prime contractor DDN’s EXAScaler Lustre-based parallel file system SW.After an 18 month evaluation the UNSW will use this integrated approach to replace 11 legacy storage systems with a coordinated environment for high-performance storage and enterprise data management. \n\nArcitecta developed an EXAScaler connector that captures file-system metadata changes for Mediaflux to use, linking high-performance storage with tools for organizing and managing data. The architecture is designed to automatically move data between high-speed on-campus storage caches and a central off-campus system. Researchers will benefit from fast access where needed, while the university can use lower-cost hard disk drives for data that doesn’t require flash-level performance. The on-premises solution will also support cloud environments.\n\nThe project began September 1, 2026. Equipment delivery is planned for early 2027, with the system expected to go live later in 2027.\n\n…\n\nJohn McKillop has joined **Hammerspace** as VP of Americas Sales. He has extensive enterprise and AI technology leadership experience, including senior executive positions at Quantum Corporation. Tom Whaley was previously VP/Head of Americas Sales at Hammerspace, leaving in May to join NetApp. \"We’ve experienced phenomenal growth this year,\" said Tony Asaro, Chief Business Officer of Hammerspace. \"We are excited to bring John onto the team to help us increase our momentum in the market. He has a great track record of success, scaling sales organizations and building the partnerships needed to turn market leadership into sustained revenue growth.”\n\nMcKillop said: \"The market opportunity is enormous, and Hammerspace is uniquely positioned to capture it. My focus is on accelerating that growth by scaling our sales organization, expanding our strategic partnerships and helping more customers realize the value of AI anywhere.”\n\n…\n\n**Hitachi Vantara** is piloting one of the world’s largest Phase Change Material installations at its Netherlands data centre with Dutch thermal energy specialist Duraflow, freeing up power for AI by cutting the electricity needed for cooling. Duraflow develops technology that absorbs, stores and releases heat, reducing reliance  on energy-intensive refrigeration. The deployment: \n\n- Delivers 240 kW of cooling using just 5–8 kW of electricity, compared with an estimated 60–80 kW for conventional cooling.\n- Uses technology that can reduce electricity consumption for thermal management by up to 90%, depending on the application.\n- Requires no water for cooling, helping support more computing capacity within the existing power connection.\n\n…\n\nSupermicro has just released a Solution Brief about a system combining **IBM Storage Ceph** with Supermicro's GrandTwin server systems, leveraging NVMe storage, 100GbE networking, and Intel Xeon 6740E processors to deliver a scalable, software-defined storage platform. More information [here](https://www.supermicro.com/solutions/Solution_Brief_GrandTwin_IBM_Storage_CEPH.pdf).\n\n…\n\nAt IBM TechXchange 2026, IBM is [announcing](https://community.ibm.com/community/user/blogs/sandra-teresa-pelaez-angeles2/2026/10/23/detect-data-theft-before-ransomware-strikes) new capabilities in the FlashSystem suite following a portfolio refresh earlier this year. FlashSystem’s ransomware detection already provides alerts within a minute of encryption. Now, IBM FlashSystem 5600, 7600 and 9600 can alert IT administrators when the data theft starts, before attackers begin encrypting the data and hold it for ransom.\n\nWhen IBM FlashSystem detects signs of potential data theft, it runs validation tests designed to confirm suspected data exfiltration in under 30 minutes, helping security teams act on high-confidence alerts while maintaining a false-positive rate below 1%. Attackers typically spend 11-14 days inside a network before they deploy ransomware, but with FlashSystem’s data exfiltration detection, the alert is raised days ahead of a typical encryption-based attack so security teams can take action earlier.\n\nIn February, IBM [introduced FlashSystem.ai](https://newsroom.ibm.com/next-generation-ibm-flashsystem-portfolio) as part of the new FlashSystem portfolio, putting AI agents to work running storage. With the latest updates announced today, FlashSystem.ai adds advice to make the agent a true partner, so routine provisioning can take up to 90% less effort. Admins can ask FlashSystem.ai questions and get answers in context with recommendations on next best actions to take, while people still approve the decisions that matter within policies they set.\n\nFor more details on these new capabilities, [**read the latest blog**](https://community.ibm.com/community/user/blogs/sandra-teresa-pelaez-angeles2/2026/10/23/detect-data-theft-before-ransomware-strikes) from IBM Storage.\n\n…\n\nDanish startup **KineDB** has raised an €850K pre-seed round, led by Superhero Capital, to build a database that handles complex data infrastructure for software engineering teams. It’s a general-purpose database designed to run wherever software runs, be it cloud infrastructure or directly on local, offline devices. KineDB initially targets sectors that rely on databases but often operate without a reliable online connection. For example, a mine worker may need to report progress or machine health via a tablet or smartphone while deep underground. With KineDB, they could update a locally stored database without an internet connection, then sync it with the wider database once the miner reestablishes a connection. \n\nIt says building software has never been so simple; AI has dramatically lowered the barrier to entry for coders. We understand both Android and iPhone target devices are actively being worked, not live yet, but available very soon.\n\n…\n\nKioxia [says](https://www.kioxia.com/en-jp/rd/technology/topics/topics-96.html) that, with the rapid growth of data driven by generative AI, 3D flash memory is increasingly required to provide higher capacity and density. One promising approach is to increase the number of stacked word lines. However, as the number of stacked word lines increases, key challenges such as cell current degradation, increased wafer warpage, and larger block size become more severe, making it difficult to realize more than 1,000 stacked word lines.\n\nTo address these challenges, a Multi-Stacked cell Array CMOS directly Bonded to Array (MSA-CBA) architecture was developed. As shown, the architecture sequentially stacks two or more memory cell array wafers on a CMOS wafer using wafer-to-wafer Cu direct bonding. By limiting the number of stacked word lines in each array wafer, it structurally mitigates cell current degradation and wafer warpage, even as the total number of stacked word lines increases. In addition, sharing the word line switch circuit among multiple array wafers helps reduce block size and lower power consumption.\n\nThe world’s first Quad-Level Cell (QLC) operation in the MSA-CBA architecture was demonstrated using prototype devices. The threshold voltage distributions of each programmed state were clearly separated immediately after QLC programming, and sufficient voltage margins were confirmed for both the individual arrays and the overall MSA-CBA structure.\n\n…\n\nMSP-focussed cyber-resilience supplier **N-able** is expanding its strategic partnership with security business SentinelOne to broaden cybersecurity capabilities  available beyond the endpoint detection and response (EDR) to cover an expanding attack surface, including AI, cloud, identities, and managed response alongside N-able Unified Endpoint Management, Adlumin Security Operations and Cove Data Protection. N-able  is now also able to deliver SentinelOne FedRAMP solutions.\n\n…\n\n**N-able** is launching AI-accelerated Vulnerability and Patch Management, designed to reduce patching windows from days to minutes. It helps teams identify vulnerabilities, prioritise urgent patches and apply updates, then verify that fixes have worked. AI-accelerated Vulnerability and Patch Management will see enhancements across N-able’s Unified Endpoint Management (UEM) platform that will reduce the patching window from days to minutes. A continuous, closed-loop workflow scans third-party applications for vulnerabilities, prioritizes the most urgent patches, remediates directly without switching tools, and verifies that the update is complete.\n\n…\n\nFor the first time, cloud file services supplier **Nasuni** has engaged a partner to deliver Nasuni Professional Services, and that partner is AHEAD, a global provider of AI-driven cloud, data, infrastructure, and security platforms, with a new HQ in Reading. AHEAD becomes Nasuni’s first partner for delivering architecture design, data migration, and cutover, which will provide customers with more  access to hands-on expertise for deployments and migrations.\n\n…\n\n**Pinecone** is releasing pg_pinecone, an open-source Postgres extension under Apache 2.0, built on the search engine Pinecone runs at scale. It indexes the vector columns you already have. Rather than a graph, its index stores a compact summary of each vector, so search stays fast on a small memory footprint and indexes build quickly. It applies filter during search, not after, so a narrow scope still returns a full set of results. It runs inside your Postgres database, so the data never leaves Postgres. The result is fast, accurate, filtered search on data you already store, with no second system to run. A 10-page [blog](https://www.pinecone.io/blog/introducing-pg-pinecone/https://www.pinecone.io/blog/) (!!!) will tell you more.\n\n…\n\nVector database and search supplier **Qdrant** says Qdrant 1.20 makes it simpler to combine keyword and semantic search on the text stored in Qdrant, with no separate keyword embeddings to generate and no statistics to maintain by hand. Qdrant 1.20 ranks text w/BM25 directly on the full-text index & derives statistics from the data itself. Keyword ranking becomes one more primitive engineers can combine with dense retrieval in a single query, with no sparse encoding pipeline to maintain. \n\nSemantic search returns results that mean the same thing as a query, but it can blur terms that must match exactly: a product code, an error number, a clause in a contract. When search misses those, a shopper gets a lookalike instead of the part they asked for, or an AI assistant answers from the wrong document. Ranking keyword matches alongside semantic ones brings those exact results to the top and keeps them ranked correctly as catalogs and document sets change.\n\nTo learn more, read the release blog [post](https://qdrant.tech/blog/qdrant-1.20.x/).\n\n…\n\nAt GMIF2026 Kevin Yoon, CVP & CTO of **Samsung Memory China** at Samsung Electronics, presented the -   - [PM1763](https://www.blocksandfiles.com/data-protection/2026/07/09/storage-news-ticker-9-july-2026/5269478) SSD, the BM1773, gen 10 BV-NAND, and zNAND-O. The BM1773 is new and a capacity flagship, built on V9 QLC, spans 16TB to 256TB and is billed as \"the world's smallest 256TB SSD.”\n\nGen 10 is the industry's first V-NAND with over 400 layers. Compared with V9, it delivers 58% higher bit density, 33% faster I/O, and 25% lower power consumption. It uses advanced bonding technology: memory cells and peripheral circuit wafers are manufactured separately and then bonded vertically, enabling high-density stacking.\n\nzNAND-O (think a high-bandwidth flash type technology) is a next-generation NAND architecture designed for on-device AI, now in development in 4-layer and 8-layer versions. It is co-packaged with the MPU and connected through the UCIe interface, delivering 10x the bit density of DRAM and 7x the read bandwidth and read energy efficiency of NAND flash. In a simulation using Moonshot AI's Kimi K2-1T model, an NPU-optimized architecture with zNAND-O cut memory cost to 1/4 at equal performance and fit a model nearly twice as large.\n\n…\n\n**Seagate** says its Mozaic HAMR disk drives “are the first and only SMR drives added to Broadcom's approved vendor list for its host bus adapter (HBA), which connects servers to storage and enables data centers to deploy and manage large fleets of hard drives.” This appears to be not quite the whole picture as a Broadcom  [Tech Doc](https://techdocs.broadcom.com/content/dam/broadcom/techdocs/us/en/dita/ca-enterprise-software/other/interop/generated-pdfs/host-bus-adapter-96xx-storage-adapter-drive-compatibility-report-ditamap.pdf) lists both WD and Seagate compatible disk drives on page 38 onwards;\n\nWe have asked Seagate to reconcile this apparent contradiction.\n\nSeagate added that “Microsoft Azure Storage has completed qualification of the Broadcom HBA with Seagate HAMR + SMR technology, providing a real-world proof point for the configuration at hyperscale.”\n\n…\n\n**Starburst** announced limited GA of [Starburst Torin](https://www.starburst.io/gpu-engine/), a GPU-native, Trino-based analytics engine for [Starburst Enterprise](http://starburst.io/starburst-enterprise). Built on Nvidia cuDF, Torin accelerates SQL workloads with native GPU execution, delivering up to 7x faster query performance.1 By complementing Starburst’s existing CPU engine, Torin gives enterprises the flexibility to run workloads on CPUs or GPUs without changing existing Trino SQL queries, applications, or agents. In Starburst benchmarks, Torin ran TPC-H queries up to 7x faster on Nvidia Blackwell GPUs, with the fastest query running 11.4x faster. On ClickBench, an Nvidia L4 GPU instance delivered 3.4x the throughput of the CPU baseline at a similar hourly hardware cost.\n\nStarburst is working to integrate open Nvidia Nemotron Retriever and Kumo models for inclusion in [Starburst AIDA](http://starburst.io/aida). Both are part of the new Agent Knowledge Foundation, which provides open models and accelerated libraries within the Agent Toolkit, helping agents retrieve relevant context, generate grounded answers, analyse and predict outcomes, and optimise actions. This work will help improve the context predictions used for fraud detection, credit risk monitoring, and know-your-customer (KYC) onboarding.\n\nRead the blog from Martin Traverso, Starburst CTO and co-creator of Trino: [*AI Agents Are Outgrowing the Converged Data Platform*](https://www.starburst.io/blog/nvidia-starburst-ai-agents-are-outgrowing-the-converged-data-platform/).\n\n…\n\nEnterprise AI Infrastructure supplier Smartbird announced a partnership with **WEKA.** It combines [WEKA NeuralMesh](https://www.weka.io/product/neuralmesh) with Smartbird’s dedicated AI clusters, giving enterprises a fully managed infrastructure foundation designed to keep sensitive data in customer-controlled environments while supporting production-scale AI. Launched in June 2026, Smartbird designs, builds, and operates dedicated AI clusters for enterprises. Each cluster is delivered as a managed service, designed for the customer’s specific workloads and deployed either on-premises or in a dedicated colocation environment. \n\n…\n\nSovereign AI cloud virtualized infrastructure supplier **Zadara** has achieved Nvidia AI Cloud Ready validation for its sovereign, multi-tenant AI cloud platform. The validation evaluated Zadara’s complete operating stack, including compute, networking, storage, data platform and GPU management. Testing spanned Nvidia HGX H100 and GB200 NVL72 systems, along with Quantum InfiniBand and  Spectrum-X Ethernet networking, confirming the platform’s readiness to run production AI workloads.\n\n…", "url": "https://wpnews.pro/news/storage-news-ticker-28-october", "canonical_source": "https://www.blocksandfiles.com/flash/2026/10/28/storage-news-ticker-28-october/5305899", "published_at": "2026-10-28 13:22:40+00:00", "updated_at": "2026-10-28 14:46:10.121582+00:00", "lang": "en", "topics": ["ai-agents", "agent-protocols", "ai-infrastructure", "ai-tools"], "entities": ["Adeptia", "Automate 5.3", "Chelsio", "CloudCasa", "Catalogic", "Dell PowerStore", "Databricks", "ActiveViam"], "also_reported_by": [], "alternates": {"html": "https://wpnews.pro/news/storage-news-ticker-28-october", "markdown": "https://wpnews.pro/news/storage-news-ticker-28-october.md", "text": "https://wpnews.pro/news/storage-news-ticker-28-october.txt", "jsonld": "https://wpnews.pro/news/storage-news-ticker-28-october.jsonld"}}