# fabric Published articles for fabric. This is one page of public article previews, not the complete archive. Follow Next page to continue. Summaries are not the original full articles. ## Microsoft's Canada timezone troubles leave admins fixing calendars, pretending they're in Bogota DevFeed: [Microsoft's Canada timezone troubles leave admins fixing calendars, pretending they're in Bogota]() Original publisher: [Read original article]() Author: Anne Barela Published: 2026-11-06T21:16:26Z Content type: news Language: en Sources: [Adafruit Industries]() Topics: [timezone](), [Microsoft](), [microsoft 365]() Tags: [calendar](), [canada](), [computers](), [daylight-savings-time](), [fabric](), [meetings](), [microsoft](), [on-prem](), [operating-system](), [outlook](), [pages](), [power-automate](), [powershell](), [reddit](), [schedules](), [sysadmin](), [time-zones](), [timezone](), [users](), [workaround]() ### AI overview After affected parts of British Columbia, Alberta, and Manitoba stopped changing clocks on November 1, some Microsoft services continued to produce wrong times or broken schedules. In Outlook, switching to the new Canadian time zone does not update existing calendar entries, so later meetings can be an hour off. Only the organizer can fix a shared meeting; changing a recurring meeting's zone resets occurrences that were moved, which must then be recreated. Microsoft says a PowerShell cmdlet for non-recurring meetings is coming, while a recurring-meeting fix remains in development. Temporary workarounds in Power Automate and Fabric use other places with matching UTC offsets: Saskatchewan, Arizona, or Bogota for Manitoba. SharePoint On-Prem is scheduled to receive correct zones November 10, and a Power Pages fix is slated for December 12. This is based on the supplied excerpt. ### Source excerpt Carly Page at The Register reports on what happened when parts of Canada stopped changing their clocks. As of November 1, affected areas of British Columbia, Alberta, and Manitoba are permanently on UTC-07:00, UTC-06:00 and UTC-05:00. Microsoft's software hasn't entirely caught up, so some customers are seeing wrong times and broken schedules. Outlook is the [...] ## AI failures threaten incident management: Here are 4 ways to avoid them. DevFeed: [AI failures threaten incident management: Here are 4 ways to avoid them.]() Original publisher: [Read original article]() Author: João Freitas Published: 2026-11-05T17:00:00Z Content type: article Language: en Sources: [The New Stack]() Topics: [incident management](), [AI Integration Strategies](), [Digital Trust](), [Post Mortem]() Tags: [2](), [3](), [adoption](), [agent](), [agent-evaluation](), [agentic](), [agentic-systems](), [agents](), [ai](), [ai-agents](), [ai-operations](), [avoid](), [can](), [devops](), [digital](), [fabric](), [here](), [incident](), [incident-management](), [management](), [one](), [operations](), [pagerduty](), [post-contributed](), [sponsor-pagerduty](), [sponsored-post-contributed](), [woven]() ### AI overview AI agents can speed incident detection and response, but their less predictable failures--including hallucinations, drift, guardrail violations, and poor tool selection--may cascade across services. The excerpt recommends four controls: set guardrails that make agents stop, report drift, or hand off to a human; use monitoring agents to detect abnormal behavior and escalate issues, with responders able to trigger pre-approved mitigations for repeat incidents; track relevant measures such as task success, groundedness, latency, tool selection, errors, cost, and escalations, using a golden set of expected behaviors; and evaluate continuously, both near real time and through post-incident reviews. It advises starting with low-risk tasks such as triage and diagnosis, then increasing automation gradually. Human oversight remains important for complex incidents. ### Source excerpt AI is becoming woven into the fabric of digital operations and incident management. On the one hand, AI can detect, The post AI failures threaten incident management: Here are 4 ways to avoid them. appeared first on The New Stack. ## GRIDSHIELD: Verifiable Grid Feasibility and Manipulation Resistance in Privacy-Preserving Peer-to-Peer Energy Trading DevFeed: [GRIDSHIELD: Verifiable Grid Feasibility and Manipulation Resistance in Privacy-Preserving Peer-to-Peer Energy Trading]() Original publisher: [Read original article]() Author: Abdullahil Kafi, Md Zahidul Islam, Alvi Ataur Khalil Published: 2026-11-05T05:00:00Z Content type: article Language: en Sources: [arXiv: Cryptography and Security]() Topics: [P2P](), [Cryptography](), [Security & Privacy](), [Bitcoin Cash]() Tags: [adversary](), [batch](), [blockchain](), [bus](), [challenge](), [circuit](), [constraint](), [cs-cr](), [energy](), [fabric](), [fairness](), [feeder](), [grid](), [manipulation](), [markets](), [metric](), [peer](), [peer-to-peer](), [privacy](), [resistance](), [trading]() ### AI overview GRIDSHIELD aims to make peer-to-peer energy trading both private and verifiably feasible on the grid. Its order-independent batch clearing is designed to prevent a compromised operator from profiting by reordering trades. An aggregated congestion-feasibility proof (ACFP) uses Pedersen commitments to verify hidden allocations against public grid constraints, without trusted setup; a reveal-and-recompute challenge is intended to catch feasible but unfair clearings. The proposed congestion-responsibility fairness metric scores curtailment against each party's physical contribution to congestion rather than consumption. In tests across four feeders, including a 240-node feeder with real hourly loads, the system reportedly produced few-kilobyte proofs, rejected infeasible allocations accepted by two of three compared designs, and sustained about 1,064 transactions per second on Hyperledger Fabric. ### Source excerpt arXiv:2611.05816v1 Announce Type: new Abstract: Blockchain-based peer-to-peer (P2P) energy markets let neighbors trade directly, but concentrate unchecked power in the clearing operator. Most prior work addresses only part of this: network-aware clearing respects grid constraints but operates on plaintext allocations, while privacy-preserving designs hide bids yet either drop the network constraint, enforce it only inside an unauditable computation, or recover it at the cost of a per-circuit trusted setup. Neither considers what happens if an adversary captures the operator's clearing privilege to reorder trades for profit or curtail a rival. We present GRIDSHIELD, a P2P energy-trading system that clears sealed bids with an order-independent batch program, removing the welfare a compromised operator can extract by reordering trades, and proves the result grid-feasible via an aggregated congestion-feasibility proof (ACFP). Exploiting the additive homomorphism of Pedersen commitments, ACFP reconstructs each grid constraint directly from per-bus commitments, reducing feeder-wide feasibility verification over any public grid constraints to a single aggregated range proof over the hidden allocation, with a reveal-and-recompute challenge that catches feasible yet unfair clearings. We introduce congestion-responsibility fairness (CRF), a metric scoring curtailment against each party's physical responsibility for congestion rather than consumption. We evaluate GRIDSHIELD against three prior P2P energy-trading designs across four feeders, including a 240-node Midwest feeder with real hourly loads, on a live Hyperledger Fabric network benchmarked with Hyperledger Caliper. GRIDSHIELD proves an entire round feasible with a few-kilobyte proof and no trusted setup, rejects infeasible allocations two of the three compared designs accept, and sustains $\approx$1064 tx/s. ## SPARTA: A Sparse Ternary LLM Accelerator with Adaptive Attention and Unified Datapath for Long-Context Edge Inference DevFeed: [SPARTA: A Sparse Ternary LLM Accelerator with Adaptive Attention and Unified Datapath for Long-Context Edge Inference]() Original publisher: [Read original article]() Author: Yanzhen Zhou, Zhengyu Han, Qunyou Liu, Zheyu Yan, Weihong Xu, Cheng Zhuo Published: 2026-11-05T05:00:00Z Content type: article Language: en Sources: [arXiv: Artificial Intelligence]() Topics: [LLM optimization](), [quantization](), [Local AI](), [Large Language Model]() Tags: [accelerators](), [adaptive](), [attention](), [bandwidth](), [cache](), [compute](), [context](), [cs-ai](), [cs-ar](), [dataflow](), [decode](), [edge](), [edge-inference](), [efficiency](), [energy](), [engine](), [execution](), [fabric](), [full](), [inference](), [llm](), [long-context]() ### AI overview SPARTA targets long-context edge inference by combining phase-aware sparse dataflow for prefill and decode with a shared compute fabric for ternary linear operations and bit-serial attention. Its two-stage attention engine uses page-level Top-k selection to reduce K-cache accesses and token-level Top-p pruning to reduce V-cache traffic. The abstract reports 4.1-4.8x less KV traffic than full attention and 1.3-1.5x less than a state-of-the-art sparse-attention scheme, with task-dependent quality loss. Against Jetson AGX Orin, it reports 3.9x generation throughput and 8.3x end-to-end energy efficiency; against Ryzen 7 9700X, 5.0x and 81.7x, respectively. Comparisons with FIGLUT, TeLLMe, and MCBP report gains in throughput, energy efficiency, and area efficiency. ### Source excerpt arXiv:2610.34824v2 Announce Type: replace-cross Abstract: LLMs on edge devices are constrained by limited memory and bandwidth. Although 1.58-bit ternary weights greatly reduce model storage, long-context inference is increasingly bottlenecked by KV-cache traffic and attention. Existing accelerators also use separate datapaths for ternary linear layers and attention, incurring hardware and execution overhead. We present SPARTA, a sparse ternary LLM accelerator for efficient long-context edge inference. SPARTA integrates: (1) a phase-aware sparse dataflow exploiting heterogeneous sparsity across compute-bound prefill and memory-bound decode; (2) a unified compute fabric mapping ternary SpMM/SpMV and bit-serial attention onto shared processing elements to exploit zero-weight and zero-bit sparsity; and (3) an adaptive two-stage attention engine combining page-level Top-k selection to reduce K-cache accesses with token-level Top-p pruning to reduce V-cache traffic. SPARTA reduces KV traffic by 4.1-4.8x over full attention and 1.3-1.5x over a state-of-the-art sparse-attention scheme, with task-dependent quality loss. Compared with NVIDIA Jetson AGX Orin, SPARTA achieves 3.9x higher generation throughput and 8.3x higher end-to-end energy efficiency on average; versus AMD Ryzen 7 9700X, the gains are 5.0x and 81.7x. It also delivers 1.6-12.0x higher throughput, 2.0-4.2x higher energy efficiency, and 1.3-51.4x higher area efficiency than FIGLUT, TeLLMe, and MCBP. ## CVE-2026-95735 - Improper Certificate Validation in HPE Networking Fabric Composer DevFeed: [CVE-2026-95735 - Improper Certificate Validation in HPE Networking Fabric Composer]() Original publisher: [Read original article]() Author: CVEFEED.io Published: 2026-11-03T19:50:52Z Content type: news Language: en Sources: [Latest High/Critical Vulnerabilitiy Feed]() Topics: [certificates]() Tags: [2026](), [certificate](), [composer](), [credentials](), [cve](), [cve-2026-95735](), [cvss](), [exploitation](), [fabric](), [hpe](), [information](), [networking](), [products](), [remote](), [validation]() ### AI overview CVE-2026-95735 is an improper certificate validation vulnerability in HPE Networking Fabric Composer, rated 8.7 (High). The source says an unauthenticated remote attacker could obtain sensitive information, including credentials, if certain preconditions outside the attacker's control are met. It marks the issue remotely exploitable, but lists no affected products or versions. ### Source excerpt CVE ID :CVE-2026-95735 Published : Nov. 3, 2026, 7:50 p.m. | 30 minutes ago Description :An improper certificate validation vulnerability exists in HPE Networking Fabric Composer. Successful exploitation could allow an unauthenticated remote attacker, under certain preconditions outside the attacker's control, to obtain sensitive information, including credentials. Severity: 8.7 | HIGH Visit the link for more details, such as CVSS details, affected products, timeline, and more... ## Diagnose Distributed AI Workload Failures with NVIDIA Attribution DevFeed: [Diagnose Distributed AI Workload Failures with NVIDIA Attribution]() Original publisher: [Read original article]() Author: Elizabeth Goodman Published: 2026-11-02T23:39:34Z Content type: tutorial Language: en Sources: [NVIDIA Developer](), [NVIDIA Technical Blog]() Topics: [AI Infrastructure](), [distributed-training](), [NVLink](), [Monitoring](), [NCCL]() Tags: [accelerated-computing-libraries](), [ai](), [ai-workloads](), [data-center-cloud](), [developer-tools-techniques](), [diagnose](), [distributed](), [error](), [errors](), [fabric](), [gpu](), [mlops](), [nccl](), [nvidia](), [nvl72](), [scheduler]() ### AI overview NVIDIA Attribution diagnoses distributed AI workload failures by correlating workload-scoped logs, metrics, telemetry, and scheduler state with deterministic rules. The Blueprint scopes a workload, collects observations, and calls the Attribution Service, which extracts and correlates events; the Blueprint returns a structured verdict with outcome, attributed domain, affected nodes, symptoms, and collector coverage. In the example, XID 149 on node-02 is identified as the root cause of a signal-integrity failure, while downstream NCCL errors are retained as evidence. When evidence supports multiple causes, the result can mark the sub-domain ambiguous; diagnosis quality depends on telemetry completeness. The self-hosted HTTP Service runs the core engine, while the customizable Python Blueprint integrates schedulers and data sources. Its quickstart uses Docker and uv with sample evidence. Rules ship with the Service and are not uploaded dynamically through its API. ### Source excerpt In distributed AI workloads, a single error can produce a cascade of symptoms. A GPU or fabric event may surface as NCCL errors on other ranks, scheduler state... ## Microsoft's Canada Timezone Troubles Leave Admins Fixing Calendars, Pretending They're In Bogota DevFeed: [Microsoft's Canada Timezone Troubles Leave Admins Fixing Calendars, Pretending They're In Bogota]() Original publisher: [Read original article]() Author: BeauHD Published: 2026-11-02T18:00:00Z Content type: news Language: en Sources: [Slashdot]() Topics: [microsoft 365](), [jsr310]() Tags: [calendar](), [canada](), [exchange-online](), [fabric](), [meetings](), [microsoft](), [outlook](), [power-automate](), [powershell](), [schedules](), [sharepoint](), [time-zones](), [timezone]() ### AI overview After parts of Canada stopped changing clocks on November 1, Microsoft reported time-zone problems across several products. Outlook users must change their calendar time zone and update upcoming events; only meeting organizers can change meetings involving others. Changing the zone resets previously moved occurrences in recurring meetings, which must be recreated. Microsoft said a PowerShell cmdlet for non-recurring meetings was expected later that week; its recurring-meeting fix was still in development with no ETA. Temporary Power Automate and Fabric workarounds use other locations, varying by product and province. SharePoint On-Prem was slated to receive the correct zones November 10, while a Power Pages service fix for dates displayed an hour off had a December 12 milestone. ### Source excerpt Microsoft is scrambling to fix time-zone problems across Outlook, Power Automate, Fabric, SharePoint and other products after parts of Canada stopped observing seasonal clock changes on November 1. "The company has confirmed problems across a chunk of its software estate, leaving some customers facing incorrect times, broken schedules and a collection of workarounds while Redmond catches up," reports The Register. From the report: Outlook users have been left with the most homework. Switching to the new Canadian time zones won't automatically update existing calendar entries, meaning that meetings scheduled after November 1 are potentially an hour out of whack. Microsoft advises users to change their Outlook calendar time zone, then update upcoming events and meetings to use it. For meetings involving other people, only the organizer can make the change. Recurring meetings come with an additional treat: changing the time zone resets any previously moved occurrences, which then have to be recreated. Anyone with a calendar full of recurring meetings may wish to put the kettle on first. Microsoft is working on an Exchange Online PowerShell cmdlet to remediate non-recurring meetings, but according to its guidance, that's not due until later this week. A fix for recurring meetings remains "in development," with an ETA yet to be confirmed. Things get stranger with Power Automate and Fabric, where Microsoft's temporary fix involves pretending you're somewhere else entirely. Depending on the product and province, that could mean Saskatchewan, Arizona or, for Manitobans, Bogota. Because apparently time zones weren't confusing enough already. SharePoint On-Prem customers, meanwhile, have to wait until November 10 for the correct time zones to arrive, while Power Pages users could see server-formatted dates appear an hour out until a service fix lands. Microsoft's current milestone for that is December 12. Read more of this story at Slashdot. ## Cisco AI Transit connects AI environments to your SD-WAN DevFeed: [Cisco AI Transit connects AI environments to your SD-WAN]() Original publisher: [Read original article]() Author: Hugo Vliegen Published: 2026-11-02T16:00:32Z Content type: article Language: en Sources: [Security @ Cisco]() Topics: [SD-WAN](), [Cisco](), [Local AI](), [GPU]() Tags: [about](), [abstraction](), [adopting](), [agentic](), [ai](), [ai-infrastructure](), [ai-workloads](), [cisco](), [cloud](), [cloud-environments](), [enterprise-networking](), [environments](), [existing](), [extend](), [fabric](), [model](), [neocloud](), [network](), [network-security](), [policy](), [sd-wan](), [security](), [teams](), [transit](), [wan]() ### AI overview Cisco AI Transit is designed to connect workloads in NeoClouds, private and sovereign AI clouds, and enterprise AI factories to an existing Cisco SD-WAN fabric, extending its shared policy model to specialized AI environments. Building on Cloud OnRamp's public cloud connectivity, it can be deployed as a physical or virtual appliance; participating NeoClouds may offer a Cisco secure virtual router as a service. Once connected, customers can use SD-WAN application policy for AI traffic classification and apply routing, segmentation, security, performance, visibility, and geographic intent. The excerpt says to confirm environment support, provider integration, controls, configuration, and operational responsibilities before planning deployment. ### Source excerpt AI workloads are spreading across specialized cloud environments. Cisco AI Transit is designed to extend the existing SD-WAN fabric to these destinations, helping network teams preserve a common policy model. ## Cisco AI Transit connects AI environments to your SD-WAN DevFeed: [Cisco AI Transit connects AI environments to your SD-WAN]() Original publisher: [Read original article]() Author: Hugo Vliegen Published: 2026-11-02T16:00:32Z Content type: article Language: en Sources: [Cisco Blogs]() Topics: [SD-WAN](), [Cisco](), [GPU](), [Local AI]() Tags: [about](), [abstraction](), [adopting](), [agentic](), [agents](), [ai](), [ai-infrastructure](), [ai-workloads](), [cisco](), [cloud](), [enterprise-networking](), [environments](), [existing](), [extend](), [fabric](), [model](), [neocloud](), [network](), [network-security](), [policy](), [sd-wan](), [security](), [teams](), [transit](), [wan]() ### AI overview Cisco AI Transit is designed to connect workloads in NeoClouds, private and sovereign AI clouds, and enterprise AI factories to an existing Cisco SD-WAN fabric, extending its common policy model as AI runs across more environments. Building on Cloud OnRamp's public cloud connectivity, AI Transit can use a physical or virtual appliance; participating NeoClouds may offer Cisco secure virtual routing as a service. Once connected, teams can use SD-WAN application policy for AI traffic and apply routing, segmentation, security, performance, visibility, and geographic intent across the fabric. Secure Internet or private connectivity are options. The source names Armada, Cirrascale, and Latitude as participating providers, and advises confirming environment support, deployment model, connectivity, operational responsibilities, and required controls before planning adoption. ### Source excerpt AI workloads are spreading across specialized cloud environments. Cisco AI Transit is designed to extend the existing SD-WAN fabric to these destinations, helping network teams preserve a common policy model. ## Unmasking fraud networks: How graph-grounded Copilots stop P2P fraud in Microsoft Fabric DevFeed: [Unmasking fraud networks: How graph-grounded Copilots stop P2P fraud in Microsoft Fabric]() Original publisher: [Read original article]() Author: Keaton Crowe Published: 2026-10-30T15:43:30Z Content type: article Language: en Sources: [Neo4j]() Topics: [Payment Fraud](), [multi-accounting fraud](), [Neo4j](), [Graph Mining](), [Artificial Intelligence](), [graph-database](), [Graphs](), [Purview](), [Azure](), [Microsoft]() Tags: [account](), [agents](), [ai](), [ai-agent](), [algorithms](), [analytics](), [architecture](), [authentication](), [azure](), [banks](), [behavior](), [device](), [fabric](), [financial-services](), [fraud](), [fraud-detection](), [graph](), [graph-database](), [ip](), [microsoft-fabric](), [networks](), [p2p](), [transactions]() ### AI overview The excerpt describes a Microsoft Fabric and Neo4j approach to P2P fraud investigations: OneLake consolidates data in Delta tables, Neo4j turns those tables into property graphs, and graph algorithms help distinguish strong links from noisy shared infrastructure. Relational joins can treat a shared private card and a public Wi-Fi IP alike, creating many false leads. Neo4j can filter high-traffic hubs, focus on rare identifiers and payment paths, and use Graph Data Science algorithms such as Weakly Connected Components to group strongly connected accounts. In the example, accounts A-102, A-104, A-106, and A-107 are grouped with fraudulent A-101, while A-103 is excluded based only on shared coffee shop Wi-Fi. The resulting graph metrics can be written to OneLake for Fabric Data Agents and Copilots to query with network context. The excerpt also names Fabric governance features and describes enabling Neo4j through the Workload Hub or Azure Marketplace. ### Source excerpt Every second, financial networks log millions of data points across disconnected internal systems: debit transactions, peer-to-peer (P2P) transfers, device fingerprints, phone numbers, and IP addresses. But when organized crime rings strike, standard relational databases hit a wall trying to trace... Read more -> ## How to secure API keys in Microsoft Fabric Gen2 dataflows with Azure Key Vault DevFeed: [How to secure API keys in Microsoft Fabric Gen2 dataflows with Azure Key Vault]() Original publisher: [Read original article]() Author: John Kerski Published: 2026-10-30T12:00:00Z Content type: tutorial Language: en Sources: [Simple Talk]() Topics: [Secrets Management](), [API keys](), [secret rotation](), [Azure](), [API](), [Microsoft]() Tags: [api](), [api-keys](), [azure](), [bi](), [cloud](), [data-analysis](), [data-analytics](), [database](), [fabric](), [featured](), [how-to](), [microsoft-fabric](), [power-bi](), [powerbi]() ### AI overview The guide describes a Microsoft Fabric Gen2 pattern for keeping API secrets out of dataflow code and source control: store the secret in Azure Key Vault, grant the workspace managed identity the Key Vault Secrets User role, then use a Fabric pipeline to retrieve the secret and pass it into an explicitly enabled public dataflow parameter at runtime. Secure Output and Secure Input mask the secret in pipeline logs, but the value remains a plain string inside the dataflow and must not be persisted or returned. Required parameters make runs fail when the pipeline does not supply a value; optional parameters can fall back to a placeholder. The pattern requires Gen2, pipeline support, and workspace admin rights, and is presented for endpoints without a suitable native OAuth connection; it does not replace available first-party secure connectors. The guide also warns to use a placeholder before Git sync and notes that an Entra client secret used in setup is visible only at creation and expires. ### Source excerpt Learn how to pass API secrets into Microsoft Fabric Gen2 dataflows using Azure Key Vault, pipelines, and public parameters (no keys in code.)... The post How to secure API keys in Microsoft Fabric Gen2 dataflows with Azure Key Vault appeared first on Simple Talk. ## ESP RainMaker Matter Fabric Support: A Complete Solution for Branded Ecosystems DevFeed: [ESP RainMaker Matter Fabric Support: A Complete Solution for Branded Ecosystems]() Original publisher: [Read original article]() Author: John Lee Published: 2026-10-30T00:00:00Z Content type: article Language: en Sources: [Blog on Developer Portal]() Topics: [Esp Rainmaker](), [Matter]() Tags: [2](), [announcement](), [app](), [blog](), [build](), [complete](), [control](), [ecosystem](), [ecosystems](), [esp](), [esp-rainmaker](), [fabric](), [keep](), [matter](), [rainmaker](), [remote](), [remote-control](), [smart-home](), [solution](), [support](), [user](), [users](), [with]() ### AI overview ESP RainMaker pairs Matter's local interoperability with cloud features for branded ecosystems: household-isolated user accounts, Fabric lifecycle and sharing management, remote control, and Matter device parsing into a human-readable Thing Model. It can issue Node Operational Certificates and manage device identity during commissioning. Teams can use RainMaker as the cloud for Matter and non-Matter devices, or integrate Fabric management, PKI, and protocol parsing with an existing business cloud through an optional API; remote calls still require a controller in the home. The open-source React Native Home app supports BLE or QR commissioning, LAN control, and cloud-connected remote control, while the SDKs support branded apps. Thread devices can join the same Fabric through a Thread Border Router, and existing non-Matter devices may be upgraded over OTA with cloud-delivered credentials and onboarding payloads. To evaluate, install the app, flash the `wifi_matter_light` firmware to a supported ESP board, and optionally use the `matter_controller` example for remote access. ### Source excerpt Build an ecosystem that supports Matter--and still keep end users on your app, with user accounts and remote control. ## The EVPN trust challenge: How to detect and contain internal MAC hijacks DevFeed: [The EVPN trust challenge: How to detect and contain internal MAC hijacks]() Original publisher: [Read original article]() Author: Sabina Anja Published: 2026-10-27T22:32:41Z Content type: article Language: en Sources: [APNIC Blog]() Topics: [evpn](), [Routing Security](), [Ethernet](), [rate-limiting](), [observability](), [Monitoring](), [trust]() Tags: [arp](), [attack](), [bgp](), [denial-of-service](), [detection](), [dhcp](), [ethernet](), [evpn](), [fabric](), [guest-post](), [mac](), [mac-address](), [mobility](), [monitoring](), [network-operators](), [route](), [security](), [spoofing](), [tech-matters](), [traffic](), [trust](), [visibility](), [vpn]() ### AI overview EVPN MAC mobility can turn a forged source MAC into a fabric-wide forwarding change: a leaf advertises the learned address with a higher sequence number, and other PEs prefer that route until the legitimate host advertises again. The article describes interception, denial of service, control-plane churn and possible IP hijacking, and notes that duplicate-MAC safeguards vary by platform: Juniper suppresses until manually cleared, while Nokia SR OS flushes and retries after 90 minutes. Detection recommendations include exporting route-reflector updates via BMP, tracking MAC moves over longer periods, and investigating sequence jumps, unexpected sticky flags, tie-break flapping and routes that conflict with access-layer records. It emphasizes validating host identity at the edge with controls such as MAC limits, DHCP snooping, ARP/ND inspection, IEEE 802.1X and sticky entries for critical systems. ### Source excerpt Guest Post: EVPN MAC mobility enables seamless workload movement, but it can also amplify the impact of MAC spoofing. A forged MAC address can redirect traffic across the fabric through BGP, making access-layer controls, route-reflector visibility, and MAC-move monitoring critical defences. ## Auto repair for Together GPU Clusters: from GPU fault to healthy node DevFeed: [Auto repair for Together GPU Clusters: from GPU fault to healthy node]() Original publisher: [Read original article]() Author: Pavneet Ahluwalia, Dongsheng Zhang, Barbara Kazanowska Published: 2026-10-27T00:00:00Z Content type: article Language: en Sources: [Together.ai]() Topics: [Distributed Systems & Parallel Computing](), [NCCL]() Tags: [2](), [3](), [account](), [actions](), [alert](), [alerts](), [also](), [annotation](), [api](), [api-key](), [automatically](), [clusters](), [code](), [fabric](), [gpu](), [gpu-clusters](), [node](), [repair](), [run](), [works]() ### AI overview Together GPU Clusters' auto-repair pipeline uses failed health checks and supporting evidence to recommend the lightest mapped fix; `NotReady` alone never triggers repair. Detection runs inside VMs and on hosts, while policy controls which faults run automatically and how long repairs wait for workloads to become idle. Repairs proceed through cordon, wait, drain, action and rejoin, with escalation after repeated failures, evidence and capacity gates for migrations, and safeguards against repair storms. Customers can submit repairs through the API or console; node-label remediation contracts are in preview. After migration or removal, the vacated machine is automatically tested and returned to service, ticketed after explicit failure, or restored to its prior state if validation is aborted. Reboots preserve local scratch data; quick reprovisioning loses it. The excerpt also describes host power cycling as a preview feature. ### Source excerpt Together GPU Clusters auto repair detects GPU, host and fabric faults, picks the lightest fix, and lets you decide per fault and per Xid code which repairs run automatically. Here is how it works. ## Singapore Boffins Weave a GUI Into Fabric DevFeed: [Singapore Boffins Weave a GUI Into Fabric]() Original publisher: [Read original article]() Author: BeauHD Published: 2026-10-23T23:00:00Z Content type: news Language: en Sources: [Slashdot]() Topics: [human computer interaction](), [User interface design](), [Embedded Systems]() Tags: [batteries](), [chips](), [compute](), [computer](), [conductive-polymer](), [controllers](), [copper](), [cycles](), [development](), [devices](), [electronic](), [electronics](), [engineer](), [external](), [fabric](), [fiber](), [flexible](), [future](), [gui](), [singapore]() ### AI overview Researchers at the National University of Singapore report a machine-washable electronic fabric that lets users interact with computing devices through clothing. Their process uses lasers to expose sections of thin copper wires and connect thousands of light-emitting chips; the wires are enclosed in transparent plastic tubing, with conductive polymer strips adding pressure sensitivity. The fibers can reportedly be made up to 10 meters long and woven into or attached to fabric. Demonstrated peripherals include a microphone, speaker, temperature sensor, and light sensor. The prototype reportedly withstood 50 gentle machine washes and 1,000 bending cycles, but still depends on external batteries and controllers. ### Source excerpt Researchers at the National University of Singapore have developed a machine-washable electronic fabric that can turn clothing into an interactive graphical user interface. The prototype reportedly survived 50 gentle machine washes and 1,000 bending cycles, though it still relies on external batteries and controllers. The Register reports: The development could lead - at some point - to a "future in which clothing functions as a pervasive, multimodal human-computer interface," according to the study published in peer-reviewed science journal Nature this week. The paper shows someone interacting with computing devices using the sleeve of a wool sweater (no jokes about the Windows ugly sweater please, this is serious). The researchers, led by Singapore National University research engineer Fu Heng, claim this is a big leap forward. "This fiber-centric approach shifts wearable electronics from flexible panels toward truly textile-native architectures, laying the groundwork for fabrics that visualize, sense, communicate and compute as naturally as the garments they inhabit," the paper says. The researchers claim to have developed a scalable process to make millimeter-thick fibers. Using lasers to strip the insulation from specific spots on the thin copper wires, they attached light-emitting chips to the exposed sections, creating electrical connections to the chips. Thousands of chips were connected to the wires with extreme precision using this process. The fibers were then fed into a transparent plastic tube, and made pressure-sensitive by adding two strips of a compressible, conductive polymer into a second longitudinal channel in the tube. The paper says they could make fibers 10 meters long using this method, and that they are thin and flexible enough to be woven into or attached to fabrics. The researchers were also able to add "peripherals" such as a microphone, a speaker, a temperature sensor and a light sensor. Read more of this story at Slashdot. ## APC Says Let There Be Light, With Raft Of Standards To Speed Up Photonics Manufacturing DevFeed: [APC Says Let There Be Light, With Raft Of Standards To Speed Up Photonics Manufacturing]() Original publisher: [Read original article]() Author: Joe Fay Published: 2026-10-23T11:46:00Z Content type: news Language: en Sources: [The Next Platform: In-depth coverage of high end computing]() Topics: [AI Infrastructure](), [Spectrum Ethernet]() Tags: [advanced-photonics-coalition](), [ai](), [ai-industry](), [apc](), [circuit](), [co-packaged-optics](), [connect](), [copper](), [cpo](), [ecosystem](), [efficiency](), [energy](), [energy-efficiency](), [fabric](), [fast](), [fiber](), [manufacturing](), [photonics](), [standards]() ### AI overview The Advanced Photonics Coalition has released five standards intended to reduce manufacturing and implementation differences, broaden supplier participation, and help scale optical networking components amid demand and supply concerns. The portfolio covers co-packaged optics architecture and practices, silicon photonics testing and manufacturing, and multicore fiber specifications. The article says APC views open standards as a way to avoid proprietary fragmentation and customer lock-in. Resolight, an APC working-group participant, says its nanosecond optical circuit switching could route data in the optical domain, reducing optical-to-electrical conversions and potentially lowering power use and transceiver needs. APC says its goal is to define a control-plane, optical-interface, and behavioral framework for fast, transparent switching, not to standardize a particular implementation. ### Source excerpt The Advanced Photonics Coalition has given the green light for a slew of standards it says should al ... ## Using AI to re-architect the Cisco DevNet sandbox environment - Part 1 DevFeed: [Using AI to re-architect the Cisco DevNet sandbox environment - Part 1]() Original publisher: [Read original article]() Author: Adrian Iliesiu Published: 2026-10-22T19:00:34Z Content type: article Language: en Sources: [Cisco Blogs]() Topics: [IP routing](), [Artificial Intelligence]() Tags: [25g](), [400g](), [agentic-ai](), [aggregation](), [ai](), [ai-agent](), [ai-for-networking](), [architect](), [architecture](), [bgp](), [bgp-evpn](), [border](), [cabling](), [cisco](), [cisco-devnet](), [cisco-nexus](), [design](), [devnet](), [devnet-sandbox](), [fabric](), [network](), [network-design](), [routers](), [sandbox](), [vpc]() ### AI overview Claude's topology review led the author to replace a planned spine-leaf fabric with a two-tier vPC design and keep two routers as spares. It first flagged that the routers should connect through border leaves because spines in the described design connect only to leaves; direct router-to-spine links also mismatched the C8475-G2's 25G ports with the 93600's QSFP ports. After clarifying that the routers would not run BGP and that classic VLANs had to span multiple leaves, the author concluded the fabric needed a two-tier vPC design, not a routed-access Clos: classic VLANs would not span a routed spine-leaf fabric without an overlay. The article also says Nexus switches can route inter-VLAN traffic in hardware at line rate, while the 8400 Series routers top out at roughly 91 Gbps, making them a poor place for fabric gateways. The author recommends sharing the full design for review, checking hardware claims, and answering requirements questions precisely. ### Source excerpt An AI agent reviewed our sandbox network design, caught routers hanging off the spines, and asked one question that turned a Clos fabric into a two-tier vPC design. ## 5 Lessons from Building a Multi-plane Network Fabric for Agentic AI DevFeed: [5 Lessons from Building a Multi-plane Network Fabric for Agentic AI]() Original publisher: [Read original article]() Author: CoreWeave Published: 2026-10-22T16:27:39Z Content type: article Language: en Sources: [Coreweave Blog Posts]() Topics: [AI Networking](), [AI Infrastructure](), [Agentic AI](), [Multi Agent Systems](), [NCCL](), [Vera Rubin NVL72]() Tags: [access-controls](), [access-management](), [adaptive](), [agentic](), [agentic-ai](), [ai](), [balancing](), [bluefield-3](), [bluefield-4](), [cks](), [compliance](), [concurrently](), [congestion-control](), [connectx](), [content-type](), [data-centers](), [fabric](), [inference](), [network](), [scale](), [training]() ### AI overview Agent workloads create unpredictable, serial chains of network hops whose latency adds up, so a fabric must carry them alongside training and inference without mutual degradation. The article describes CoreWeave's approach: 1:1 non-oversubscribed links; Spectrum-X adaptive routing and telemetry-based congestion control; multi-plane and multi-rail designs to scale while keeping paths short; and ConnectX-9 hardware load balancing that reacts to congestion without host CPU scheduling. It also describes resilience across topology, BlueField DPUs, software, operations, and geography, plus flow-level observability and Telemetry Relay. These are the provider's reported design choices and capabilities; the text offers no independent comparative measurements. It notes that congestion controls require workload-specific tuning, since aggressive pausing can cause head-of-line blocking and tail latency. ### Source excerpt Five lessons we learned from building a multi-plane network that can handle training, inference, and agentic workloads concurrently, at scale. ## A Blank Slate in Buffalo: How the Bills Built the NFL's First Complete Wi-Fi 7 Stadium DevFeed: [A Blank Slate in Buffalo: How the Bills Built the NFL's First Complete Wi-Fi 7 Stadium]() Original publisher: [Read original article]() Author: Shaun Handley Published: 2026-10-21T19:03:13Z Content type: article Language: en Sources: [Cisco Blogs]() Topics: [network security](), [Wi-Fi](), [observability](), [it modernization]() Tags: [access-points](), [architecture](), [broadcast](), [capacity](), [catalyst](), [cisco](), [cisco-data-center](), [cisco-secure-firewall](), [cisco-secure-networking](), [cisco-service-provider](), [cisco-ucs](), [compute](), [concurrent](), [connectivity](), [data](), [deployment](), [devices](), [fabric](), [firewall](), [industries](), [network](), [security](), [wi-fi](), [wi-fi-7]() ### AI overview The Bills replaced Highmark Stadium's aging, multi-vendor network with a unified Cisco architecture, built and prepared in a separate tech building before the stadium was finished. It combines Catalyst switching, Wi-Fi 7 access points, Cisco IPFM with N9000 switches, embedded Cisco Secure Firewall, and operational tools including Cisco Spaces and Stadium Vision, with UCS compute and Nutanix storage. The article says integrating security into the same fabric reduces the separate systems staff must monitor and reconcile. The network was engineered for 80,000 concurrent devices in a 60,000-seat stadium, and the Bills report improved Voice of the Fan ratings. The author calls it the NFL's first complete Wi-Fi 7 deployment; the text provides no independent validation or comparative performance data. ### Source excerpt Shaun Handley, Director of Stadium Technology for the Buffalo Bills, on building Highmark Stadium's network and security from a blank slate. ## NVIDIA Spectrum-X: Hardware-Accelerated Multiplane Networking for Large-Scale AI Training DevFeed: [NVIDIA Spectrum-X: Hardware-Accelerated Multiplane Networking for Large-Scale AI Training]() Original publisher: [Read original article]() Author: Sajy Khashab, Albert Gran Alcoz, Alon Gal, Jacky Romano, Rani Abboud, Yonatan Piasetzky, Lior Maman, Amit Nishry, Barak Gafni, Omer Shabtai, Matty Kadosh, Dror Goldenberg, Gilad Shainer, Mark Silberst Published: 2026-10-21T04:00:00Z Content type: article Language: en Sources: [arXiv: Distributed Computing](), [arXiv: Networking]() Topics: [NVIDIA Spectrum-X Ethernet](), [AI Networking](), [distributed-training]() Tags: [ai](), [ai-infrastructure](), [ai-training](), [architecture](), [balancing](), [bandwidth](), [benchmarks](), [concurrent](), [cs-ai](), [cs-dc](), [cs-ni](), [distributed](), [efficiency](), [ethernet](), [evaluation](), [fabric](), [hardware](), [isolation](), [latency](), [llm](), [networking](), [nvidia](), [spectrum-x]() ### AI overview For large-scale AI training, NVIDIA Spectrum-X uses a multiplane Ethernet architecture that substitutes topological parallelism for hierarchical network depth, with NIC and switch hardware accelerating load balancing to respond to changing conditions at microsecond timescales. The abstract reports 98% of theoretical line rate with low, jitter-free latency, cross-tenant isolation, capacity-proportional bisection bandwidth, and a 7% latency increase under 10% fabric link failures. It also reports rapid reaction to host and fabric link flaps during LLM training. The supplied excerpt does not provide detailed benchmark methods or deployment lessons. ### Source excerpt arXiv:2605.21187v2 Announce Type: replace-cross Abstract: As distributed model training scales to span hundreds of thousands of GPUs, scale-out networks face unprecedented performance and efficiency demands. NVIDIA Spectrum-X Ethernet has been designed from the ground up to achieve predictable and stable network performance with high utilization and low latency. This paper presents the Spectrum-X multiplane architecture, which replaces hierarchical depth with topological parallelism, and introduces hardware-accelerated load balancing in NICs and switches as the key architectural approach to provide fast reaction to highly dynamic network conditions at the microsecond timescales that AI training workloads demand. We describe the motivation, design principles, evaluation methodology and performance on state-of-the-art benchmarks, as well as the lessons we learned from deploying and debugging Spectrum-X networks in large-scale systems. Our evaluation highlights production-grade AI infrastructure performance across four core dimensions: 98% of the theoretical line rate with low jitter-free latency; strong cross-tenant isolation for concurrent workloads; robust, capacity-proportional bisection bandwidth and 7% latency increase for 10% fabric link failures; and rapid reaction to host and fabric link flaps during LLM training workloads. ## Microsoft Commits Maia to Multi-Gigawatt Inference, and Its Efficiency Metric Draws Questions DevFeed: [Microsoft Commits Maia to Multi-Gigawatt Inference, and Its Efficiency Metric Draws Questions]() Original publisher: [Read original article]() Author: Steef-Jan Wiggers Published: 2026-10-16T06:02:00Z Content type: news Language: en Sources: [InfoQ]() Topics: [Distributed Systems & Parallel Computing](), [NVLink](), [Latency]() Tags: [accelerators](), [ai](), [ai-ml-data-engineering](), [architecture](), [architecture-design](), [azure-maia-300](), [cloud](), [development](), [devops](), [ethernet](), [fabric](), [hardware](), [inference](), [maia](), [maia-300](), [metric](), [microsoft](), [ml-data-engineering](), [news](), [performance](), [scale](), [scale-out](), [scale-up](), [tokens-per-watt]() ### AI overview Microsoft plans multi-gigawatt deployments of its Maia inference accelerators from 2027, with Maia 300 using a unified Ethernet fabric to connect accelerators and move KV state between prefill and decode workers. The article describes the architecture and reports questions about Microsoft's tokens-per-watt efficiency claims, including their measurement and applicability to customer workloads. ### Source excerpt Microsoft will scale its Maia accelerator family to multi-gigawatt deployments from 2027, making it the primary inference system for its own products. Maia 300 collapses scale-up and scale-out onto one Ethernet fabric and moves KV state between prefill and decode workers without leaving it. Practitioners have questioned the tokens-per-watt claim that leads the announcement. By Steef-Jan Wiggers ## Building a Governed Customer 360 Data Product on Microsoft Fabric DevFeed: [Building a Governed Customer 360 Data Product on Microsoft Fabric]() Original publisher: [Read original article]() Author: Saradha Published: 2026-10-16T04:59:25Z Content type: tutorial Language: en Sources: [HackerNoon]() Topics: [Purview](), [AI-Ready Data](), [data-engineering](), [AI Infrastructure](), [Data Space](), [implementation](), [Embeddings](), [AI Chat]() Tags: [ai](), [ai-ready-data-products](), [customer-360](), [customer-360-data-platform](), [data-governance](), [enterprise-data-architecture](), [fabric](), [fabric-iq](), [how-to](), [mcp](), [mcp-tools](), [microsoft](), [microsoft-fabric](), [microsoft-onelake](), [models](), [power-bi-semantic-models](), [semantic](), [semantic-models](), [tools]() ### AI overview This tutorial shows how to build a governed Customer 360 data product on Microsoft Fabric, from ingestion and data quality checks to semantic context, access controls, and MCP tools for AI agents. It emphasizes reusable data contracts and constrained, auditable access; some Fabric IQ and MCP features are in preview. ### Source excerpt Learn how to build an AI-ready Customer 360 data product with Microsoft Fabric, governed data layers, semantic models, and MCP tools. ## This is the AMD Helios HPE Juniper QFX5252 Switch Tray at OCP 2026 DevFeed: [This is the AMD Helios HPE Juniper QFX5252 Switch Tray at OCP 2026]() Original publisher: [Read original article]() Author: Patrick Kennedy Published: 2026-10-16T04:49:49Z Content type: article Language: en Sources: [ServeTheHome]() Topics: [Cloud Networking](), [radeon-memory-visualizer](), [Ubiquiti UCG-Fiber]() Tags: [2026](), [amd](), [broadcom](), [copper](), [cpo](), [dpu](), [fabric](), [helios](), [high-density](), [hpe](), [juniper](), [liquid-cooling](), [networking](), [ocp](), [switch](), [tomahawk]() ### AI overview The article examines the AMD Helios rack's HPE Juniper QFX5252 switch tray, describing its paired Broadcom Tomahawk 6 switches, liquid cooling, retimers and dense copper cabling. It explains how the design supports scaling up within a rack and contrasts the cable-heavy setup with the potential simplification of co-packaged optics. ### Source excerpt At OCP 2026, we saw an AMD Helios AI rack switch tray, the HPE Juniper QFX5252. This just looks super cool The post This is the AMD Helios HPE Juniper QFX5252 Switch Tray at OCP 2026 appeared first on ServeTheHome. ## Cosmos DB Mirroring in Fabric Now Supports Virtual Network Data Gateways DevFeed: [Cosmos DB Mirroring in Fabric Now Supports Virtual Network Data Gateways]() Original publisher: [Read original article]() Author: Jai Maldonado Published: 2026-09-30T07:00:35Z Content type: release Language: en Sources: [Azure Cosmos DB Blog]() Topics: [Amazon VPC](), [network security](), [Cloud](), [Infrastructure as a Service (IaaS)]() Tags: [analytics](), [announcements](), [azure-cosmos-db](), [azure-cosmos-db-for-nosql](), [fabric](), [microsoft-fabric](), [network-configuration](), [network-security](), [networks]() ### AI overview Microsoft Fabric's Azure Cosmos DB mirroring now supports virtual network data gateways for accounts protected by private endpoints or virtual networks. The gateway keeps public network access disabled during setup and replication, while eliminating the need to configure Microsoft service tag IP ranges. ### Source excerpt Azure Cosmos DB mirroring in Microsoft Fabric now offers a simpler way to connect accounts secured by private endpoints or virtual networks. A virtual network data gateway provisioned inside your own virtual network keeps public network access disabled throughout setup and ongoing replication. This release builds on the general availability of private network support announced [...] The post Cosmos DB Mirroring in Fabric Now Supports Virtual Network Data Gateways appeared first on Azure Cosmos DB Blog. [Next page]()