Resources
Insights on mainframe operations
Notes on running z/OS operations, from the people who build Meridian. Plain words, no slides, and nothing we would not say to an operations manager at three in the morning.
-
Fortinet turns firewall licences into an AI security on-ramp
Fortinet launches a cloud-delivered AI security service that runs on existing FortiGate firewalls, giving enterprises fast, low-cost AI protection.
-
Zero-Trust Strategies Often Miss Mainframe Workloads, Creating Risk for Enterprises
Enterprise zero-trust programmes frequently overlook mainframe environments, leaving core transaction systems exposed to attack through integration gaps.
-
Upwind secures $400 million funding as AI-driven cloud security gains traction
Upwind Security raised $400 m at a $3.8 bn valuation, reporting revenue quadruple in 2026 and highlighting strong market appetite for AI cloud security.
-
Identity-based safeguards limit misbehaving AI agents in corporate environments
Okta outlines how giving each AI agent a unique identity and scoped authorisation stops unintended actions and improves audit trails for enterprises.
-
Oracle integrates Google TPU accelerators into Oracle Cloud, expands AI compute options
Oracle will offer Google’s TPU chips in its public cloud from next year, adding new AI hardware and ChatGPT-based services for enterprise customers.
-
Red Hat integrates OpenShell to secure enterprise AI agents ahead of KubeCon
Red Hat announces OpenShell, an open runtime that isolates AI agents like OpenClaw, aiming to protect enterprise deployments from emerging threats.
-
Cisco launches task-scoped privileges for admins and AI agents on network devices
Cisco's Privileged Infrastructure Access ties admin and AI rights to specific tasks, cutting excess access and improving audit trails.
-
How Enterprises Transition from AI Pilots to Autonomous Operations
Enterprises must evolve from AI insights to trusted autonomous actions, balancing automation, security, and network readiness as AI agents assume control.
-
Hybrid infrastructure choices are reshaping AI cost models for large enterprises
Enterprises move AI inference to hybrid setups as token costs rise, prompting new spending strategies and edge deployment.
-
Okta details new AI agent governance controls at Oktane
Okta’s Oktane event highlighted shared architecture and cross-vendor cooperation to govern AI agents in identity workflows and prevent misuse.
-
What clients expect from a managed batch‑monitoring service
Clients of managed service providers need clear reports, reliable SLAs, audit‑ready evidence and rapid response for z/OS batch workloads.
-
Google Cloud unveils Agent Compute Engine to run AI agents safely and efficiently
Google Cloud's Agent Compute Engine uses microVM sandboxes, Titan chip firmware checks and idle detection to cut power use and boost AI agent.
-
AI agents develop new skills when left to play on a simulated island
MIT researchers let AI agents explore a virtual island without goals and found they learned useful behaviours, showing play can teach machines.
-
OpenAI makes its newest GPT-6 models available via Google Cloud
OpenAI will host GPT-6 Astra and smaller variants in Google Cloud’s Model Garden, giving enterprises direct access to advanced, safety-enhanced LLMs.
-
AI agents inflate telemetry spend and add network observability hurdles
Research shows AI agents raise telemetry costs, cause blind spots and demand new monitoring practices for enterprise networks.
-
Quick machine recovery now supported on Windows Server 2025 after October 2026 update
Microsoft's October 2026 update adds quick machine recovery to Windows Server 2025, letting admins fix boot failures remotely and reduce downtime.
-
doubleAI secures $200 million Series A as AI optimisation gains focus
doubleAI’s $200 million Series A, led by Lightspeed and Bessemer with Nvidia and Dell, funds its Maxwell model that cuts AI hardware use.
-
TorchNeuron lets enterprises run PyTorch on AWS Trainium with minimal code changes
Amazon's open-source TorchNeuron toolkit enables PyTorch workloads on Trainium2 and Trainium3 chips, cutting AI training costs for enterprise teams.
-
Drupal releases patches for two critical remote code execution vulnerabilities
Drupal has issued patches for critical RCE flaws in its core Field UI and Entity List modules, urging admins to update before October 14.
-
Claude Mythos model sparks mainframe security reassessment over TN3270 emulators
Anthropic's Claude Mythos uncovers decades-old flaws, prompting mainframe admins to rethink TN3270 security and AI-driven tooling.
-
How to build a statutory compliance calendar for Indian entities
Learn how to list PF, ESI, professional tax and LWF returns, set entity‑specific due dates and create reminder rules to avoid missed deadlines.
-
Investors put money into AI agent security as agents outrun enterprise controls
Pluto Security raised $42 million and Reco $55 million to find and police AI agents at work. What the money says, and what enterprise IT teams should do.
-
Rocket Software buys Geniez AI as AI agents reach the mainframe
Rocket Software has bought Geniez AI, which builds AI agents for the mainframe. What z/OS teams should check while the nightly batch still has to finish.
-
Sompo Japan rehosts a mainframe system on Azure, keeping its COBOL
Sompo Japan rehosted a COBOL, JCL and assembler system on Azure after 45 months, as AWS showed a rewrite route. What it means for mainframe teams.
-
Key CA 7 queues and job states to watch each night for early batch-delay detection
Learn which CA 7 queues, job statuses, abends and prerequisites to monitor nightly so you can spot a late batch before it breaches its SLA.
-
Checklist for evaluating a mainframe batch monitoring tool
A practical checklist to assess batch monitoring, prediction, automation, multi‑client support, evidence and footprint for z/OS estates.
-
A z/OS operations glossary: the words on an operations floor
Plain definitions of the z/OS operations terms that come up every night: LPAR, WTOR, SMF, abend, initiator, spool, scheduler, MSU and more.
-
WTOR automation on z/OS: how to reply safely across many LPARs
What a WTOR is, why unanswered and wrongly answered WTORs hurt, and how to automate replies across many LPARs without losing control.
-
Running many clients' mainframes from one operations floor
What changes when a mainframe service provider runs twenty client estates instead of one, and what the operations tooling has to do about it.
-
Restarting a failed CA 7 job without missing the batch SLA
A failed CA 7 job at night: how to choose between restarting from a step, from the top or forcing it complete, with the batch deadline in view.
-
How to predict a batch SLA breach on z/OS, hours before it happens
How batch SLA prediction works on z/OS: the job chain, run-time history, the critical path, live subsystem status, and saying it in plain words.
See it on your own estate.
A ninety-day proof of concept on one LPAR, free, with our engineers on the call. After ninety days you decide on your own results.