Cybersecurity NEWS. THE AMIS PERSPECTIVE.

What’s new. What matters. What’s next.

Cybersecurity & Technology products, innovations, and security developments, translated into practical questions for your business. Less noise. More context.

Security & Architecture Insights.

NEWEST FIRST

Outage

Nationwide ground stop lifted after more than 4 hours as Trump administration blames old, overloaded tech

Flight delays and cancellations after nationwide ground stop

Read Full Article

AMIS TAKE

Rough week for uptime. Microsoft 365 had service issues. The FAA halted every departure in the country for more than 4 hours over “network inconsistencies.” Then AWS US-EAST-1 took down IAM Identity Center sign-in. Somewhere out there is someone who couldn’t get into Outlook, couldn’t catch their flight and couldn’t log in to their AWS account, all in the same week. Joking aside, AWS handled theirs well. Problems started at 3:05 AM PST, engineers were on it by 3:10, authentication was back by 5:20 and the last load balancer fix landed at 6:21. The impact stayed contained, they engaged fast and their updates were clear. The FAA outage is the one I keep coming back to, mostly because of the “why.” The blame is on legacy telecom: an air traffic control system that still runs on analog telecom. The message from the podium was two things at once. We have the safest airspace in the world, and without funding for a complete overhaul, this will happen again. This is the explanation that caught my attention: “Our switches, our routers, actually got overloaded with data ... and started to give us inconsistency in the data from those routers.” To me, that reads like a Layer 3 problem. It could be dynamic routing instability, a multicast routing issue, or a change rolled out somewhere that sent traffic this legacy infrastructure was never built to handle. The fix being “shut it down and bring it back” also fits an internal routing problem. That’s speculation. There’s no official root cause report yet. DHS says it doesn’t appear to be a cyberattack but isn’t ruling one out. The contractor behind the SMART AI system being rolled out nationwide says its system wasn’t the cause. The ironic part: the public’s first instinct was “cyberattack,” and the second was “cover-up.” Cyberattacks have never gotten more attention, yet the companies that protect organizations from them are watching their stock prices slide. Not every outage is an attack. But old infrastructure carrying new traffic is a risk you can see coming.

Outage

Microsoft Struggles to Restore Email and Other Services in a Global Outage

From Asia to the U.S., Microsoft’s outage involving services like Outlook and Teams has dragged on for hours and impacted users, companies and some government operations as the technology giant struggled to fix the issue.

Read Full Article

AMIS TAKE

The Microsoft 365 outage wasn't a traffic problem. It was a health-check problem. Since about 10 AM Pacific yesterday, Outlook, Teams, SharePoint and Copilot have been down or degraded for users around the world, from the US to Japan, Australia, Hong Kong and India. As of this morning, Microsoft is still working through recovery. I followed Microsoft's status updates. A few things stood out: - The first explanation was "elevated load." An hour later Microsoft walked that back: the load wasn't coming from more traffic. Servers were dropping out unexpectedly and pushing their share onto whatever was left standing. - The suspected trigger is the component that decides whether a server is healthy. Servers failed automated health checks, got pulled from rotation, and the survivors buckled. The mechanism meant to protect availability became the outage. - The fixes didn't stick. Servers brought back online went unresponsive again once traffic hit them, and by evening Microsoft said its changes weren't taking as expected. - In my own environment, Teams came back mid-day while Outlook stayed dark much longer. Email wasn't called out by name in Microsoft's updates until the afternoon. Microsoft says there's no sign of a cyberattack, and early in the day said it hadn't found a recent change that lined up with the start of impact. My question is: what changed? Servers don't start failing health checks on their own. With how fast hyperscalers are reshuffling infrastructure for AI right now, it will be interesting to know whether capacity moves or decommissioning played a part. That's a hunch, not a finding. The post-incident review will tell us. What I'm telling clients today: 1. Expect outage-themed phishing. Fake "reset your password" and "restore your mailbox" lures showed up within hours. Microsoft isn't emailing you to fix this. 2. Get status from the M365 admin center (MO1550821 / EX1550543), not from links in your inbox. 3. Have an out-of-band channel. If your incident comms run through Outlook and Teams, you don't have incident comms. 4. Know where inbound mail goes when Exchange Online can't accept it. A gateway that can queue mail is the difference between a delay and lost messages. Monitoring tells you something is down. Architecture decides what you can still do while it's down.

Vulnerabilities

Asus, Zoom Release Security Patches

Asus and Zoom are among the tech companies that have released fresh security updates for their products this week.

Read Full Article

AMIS TAKE

Zoom is the way in. Your laptop is the target. Zoom patched two flaws in Zoom Workplace for Windows this week: CVE-2026-108186 (Critical, 9.6) and CVE-2026-108187 (High, 8.1). The headlines call it "privilege escalation." That undersells it. Here's what the CVSS vector on the critical one tells you: → Reachable over the network → No credentials needed → One user interaction (a click, an invite, a message) → Scope changed: the impact doesn't stay inside Zoom That last point is the important one. "Scope changed" means exploiting the app reaches past the app, into the Windows machine it runs on. So the chain looks like this: 1. Attacker sends something crafted 2. User interacts with it 3. The Zoom client mishandles it 4. Attacker has a foothold on the endpoint Once they're on the endpoint, they inherit everything that user can reach, including the signed-in Zoom session: meeting summaries, chat history, shared files, AI notes from your strategy calls. And Zoom Workplace isn't a niche add-on. It's the Zoom app on nearly every corporate laptop. What to do: • Update to 7.1.8 or later. That covers both CVEs. • 7.1.5 through 7.1.7 are still exposed to the critical one. • Zoom lists no workaround. Patching is the only control. • If you can't see which Zoom version is on every endpoint, that's your real finding. No known exploitation yet. That's your window, so use it.

AI

The Biggest Threat to Your GRC Job Isn’t AI. It’s Ignoring AI

The question in GRC is no longer whether your job is safe. It is whether your skillset is. AI agents are already doing the control testing, the evidence collection, and the framework mapping that used to fill a manager's week, and they get better at it every quarter.

Read Full Article

AMIS TAKE

AI agents won’t replace GRC. But they will expose the habits we’ve been getting away with. I read a piece on BleepingComputer this week about the skills GRC teams need as agents take over control testing and evidence collection. Three stood out, and the more I thought about them, the clearer it got: none of them are new. They’re skills we’ve always needed and often skipped. 1. Data literacy GRC sits in an unusual spot. You need to understand the business well enough to see its risks, and the technology well enough to test a control against policy. To me, data literacy means knowing exactly where control evidence comes from and what it has to show for an accurate assessment. Many GRC leaders have leaned on analysts to interpret that evidence. Hand the work to an agent and that buffer is gone. 2. Define “good” before you deploy The article suggests treating every agent like a direct report: set the outcome, accuracy bar and escalation rate up front. Honestly, most of us don’t do that with people. We find out what good looks like after a few quarters of chasing it. AI adoption pushes for speed, but moving work to agents means slowing down first to define what we’re actually trying to improve. 3. Write it like a runbook Agents need instructions precise enough that they can’t be misread, and every action has to land somewhere a human can review it. Plenty of teams don’t get that level of clarity from leadership today, especially when leaders are busy keeping the business running or scaling it. The takeaway: agents raise the bar on management discipline. Teams that write it down, measure it and make it visible will get real value from AI. Everyone else will just automate their gaps. How is your team defining “good” before agents go live?

Tools

Top 10 Security Tools every security engineer must know

In this article, we will discuss the top 10 security tools security engineers must know concerning different areas like Pentesting, Vulnerability scanning, incident response and more.

Read Full Article

AMIS TAKE

It’s a solid list, and I agree with the main point. A tool only helps as much as you understand what it’s showing you. But most of these tools assume the network underneath is working. In my experience, across helpdesk, engineering, architecture and management, a lot of “security problems” turn out to be reachability, routing or DNS problems. Here’s where I start before I reach for the security stack: - Ping. Reachability first, every time. Continuous ping with -t, and -S to set the source IP when a host has multiple NICs. - Traceroute / tracert. My map for VPNs and internal routes. It shows the hop where packets stop when static or dynamic routes aren’t behaving. - nslookup. If ping works but “the internet is down,” check which DNS server the host is using and whether it answers correctly. - Wireshark. The one tool on both lists. A minute or two of capture can explain “weird behavior” when you don’t have logs or visibility. - PuTTY. In an outage, the CLI is often the fastest, safest and sometimes only way in. On a FortiGate with CPU or memory spiking, diagnose sys top finds the runaway process so you can stabilize the network. Security tools tell you what’s wrong. Network fundamentals tell you whether the network can carry traffic at all. You need both, and in an outage the basics usually come first.

Software

Agentic Data Security Firm Sentra Raises $150 Million

Agentic data security company Sentra today announced raising $150 million in a Series C funding round that brings the total raised by the company to over $260 million. The investment round was led by Capital One Ventures

Read Full Article

AMIS TAKE

AI security money is moving fast this week. Sentra raised a $150M Series C for what it calls agentic data security. The day before, CrowdStrike announced it's acquiring Fiddler AI and Dreadnode, reportedly around $495M combined. I compared Sentra to Fortinet's FortiAI, and they're solving different parts of the same problem: - FortiAI sits on top of the telemetry you already have. With logging, IPS, and the Fabric turned on, it correlates that data and surfaces patterns at a speed no analyst can match. It also polices AI traffic in motion: shadow AI discovery, guardrails on prompts. - Sentra starts with data at rest. Find and classify sensitive data, map what users, service accounts, and AI agents can actually reach, and pull back access that shouldn't exist. One watches what's moving. The other answers, "What could this agent get to if it went sideways?" CrowdStrike's buys cover the other angles: Fiddler watches what agents are doing, and Dreadnode attacks them before someone else does. Sentra is a focused startup going after one specific problem, which is exactly the profile that gets acquired. I wouldn't be surprised to see a larger platform pick it up. The takeaway for anyone rolling out AI agents: AI-driven visibility is only as good as the data you feed it. Turn the logging on first.

Cyberattack

Frontline education data breach impacts over 1.3 million

Frontline Education is notifying over 1.3 million people that their personal information was compromised in a data breach.

Read Full Article

AMIS TAKE

“We are only as strong as our weakest link.” That’s true in any team sport, and it’s true in IT. Frontline Education makes school administration software used by K-12 districts across the U.S. It just disclosed a breach affecting over 1.3 million people. Attackers used a vulnerability in a third-party software product Frontline uses to get into part of its environment on August 14–15. The exposed data may include Social Security numbers, driver’s license numbers and other government IDs. Frontline is giving affected U.S. adults two years of free credit monitoring and identity protection, plus monitoring for minors. Its notice also lists stronger monitoring, alerting and response as part of the fix. Right now the pressure is to integrate everything. If you aren’t connecting AI or third-party tools to your workflows, apps and systems, you’re “falling behind.” What’s easy to look away from is that every one of those tools and vendors is a way into your environment, and into your organization’s data and your customers’ data. Three controls come to mind from this one: 1. Third-party risk assessments: Before a vendor or app gets plugged in, ask what controls protect its infrastructure, business and software. Then weigh the risk it brings to you and your customers. 2. Vulnerability management: Find, analyze, respond, monitor. In short: patch what’s vulnerable, and that includes the integrations your critical systems depend on. 3. Log reviews: This is the monitoring part, and probably the most important. The controls that block attacks or send alerts only matter if someone reviews them, every hour of every day. If a tree falls in a forest and no one is around, does it make a sound? Yes. The threat actor will let everyone know… eventually. How are you vetting the third-party tools connected to your environment?

Cyberattack

Revolve Data Breach Impacts at Least 180,000 People

Online fashion retailer Revolve is notifying at least 180,000 individuals that their personal and payment card information was stolen in a data breach.

Read Full Article

AMIS TAKE

Revolve isn’t unique. Businesses like this put their energy where it belongs: building a customer base, delivering their main product, and growing. Fashion, shipping and logistics take a lot of attention, and technology often sits in the background, even when the whole business runs online. That’s a tough spot to be in. When your focus is far from technology, it’s hard to see how much cybersecurity matters until something like this happens. I’d be curious how Revolve’s IT and security teams are set up and how much support they get as the business grows. It reminds me of the Target breach, on a smaller scale and with a smaller news cycle. The lesson carries over, though: customer data comes with responsibility, and that responsibility grows with the business. A feature like saved payment cards is a good example. It makes shopping easier for customers, and it also gives the business more to protect. Customers trust businesses with their personal information every time they check out. Protecting that data is a big part of keeping that trust, and it’s something every growing business has to plan for, whatever it sells.

Vulnerability Management

Why Urgent Security Fixes Aren't Fixed. Observe, Prove, Act, Confirm Gets Them Moving

Prioritization establishes where action should begin. Safe remediation gives teams a controlled path for carrying that decision into production, with evidence supporting the proposed change and confirmation showing its result.

Read Full Article

AMIS TAKE

Reporting and monitoring are how an organization sees the value of its security team without getting hacked first and calling that team in a panic to restore systems. A simple, consistent change request turns security from the team that slows things down into the team that shows its work. If your team is small and your approval cycles are long, AMIS can help you build the reporting and monitoring that make the case for you.

Read the full AMIS take

SOFTWARE / INNOVATION

FortiEndpoint Expands Security for the AI Era

Today, we’re expanding FortiEndpoint with powerful new capabilities designed for the AI era. These latest innovations expand our unified endpoint platform by adding visibility and control, native data protection, endpoint risk management, FortiAI-assisted operations, and flexible deployment options for organizations with data sovereignty requirements. Combined with advanced endpoint protection, EDR, and universal ZTNA, FortiEndpoint delivers comprehensive endpoint security through one agent, one console, and one license.

Read Full Article

AMIS TAKE

Fortinet is betting on one agent instead of a new agent for every feature. FortiEndpoint now brings FortiClient, FortiEDR and FortiDLP together in a single agent, console and license. It covers zero-trust access, AI governance, endpoint protection, EDR, data security and insider risk, with AI-assisted threat hunting on top. That puts real pressure on standalone FortiEDR. The classic reason to buy FortiEDR on its own was the managed option, with Fortinet analysts reviewing your alerts. FortiEndpoint already offers managed tiers with incident response included, and it starts at 25 seats; standalone FortiEDR bundles start at 100 to 500. I wouldn’t be surprised to see FortiEDR fold into FortiEndpoint entirely. For now it sticks around for on-premises deployments and for existing FortiEDR customers, who can’t move to FortiEndpoint directly. So why buy anything else for endpoint protection? There’s one consideration before you buy FortiEndpoint, and it could cut what you pay for by more than you’d expect.

Read the full AMIS take

SOFTWARE / INNOVATION

Fortinet Unveils New Calgary Cybersecurity Innovation Hub

Fortinet®(NASDAQ: FTNT), the global cybersecurity leader driving the convergence of networking and security, today unveiled its cutting-edge cybersecurity Innovation Hub in downtown Calgary, Alberta. The new facility reinforces Fortinet’s long-term investment in Canada and reflects the company’s commitment to strengthening cyber resilience and advancing cybersecurity innovation, education, and customer engagement.

Read Full Article

AMIS TAKE

Fortinet put a new SASE PoP in Calgary, and it fills a real gap on the map. In Fortinet’s published FortiSASE data center list, its Canadian Fortinet Cloud locations are all in BC and Ontario. Kamloops to Toronto is about 3,150 km with nothing listed in between. The new Calgary Innovation Hub changes that. Why it matters if you run FortiSASE: - It’s Fortinet’s own infrastructure: its own appliances and ASICs, in a building it owns. It isn’t rented public cloud capacity. - Fortinet Cloud locations are listed for every FortiSASE tier. Most public cloud locations require Advanced or Comprehensive. - Users and branches on the Prairies get a closer in-country option, plus one more Canadian failover choice that doesn’t send traffic across the border. - Data sovereignty: for many Canadian organizations, keeping traffic in Canada is a requirement, not a nice-to-have. Fortinet investing in its own PoPs tells me demand for securing remote users isn’t slowing down. The part I like most isn’t the PoP. The building also has a training facility, a customer experience centre, a community hub with an OT and critical infrastructure lab, and a new merit-based award at the University of Calgary ($3,000 USD scholarship for five years). Tech isn’t all web development. When a tech company comes to a new city, showing people the full range of the field is the best thing it can do.

SOFTWARE / FORTIOS PLATFORM

3 Key Takeaways from the Fortinet SASE Summit 2026

Three key themes ran throughout the summit. First, AI is reshaping what organizations must protect and introducing autonomous agents as a new class of enterprise user. Second, IT and security teams need tighter integration across networking, security, identity, and operations. And third, new sovereignty requirements are prompting organizations to reconsider where their security services operate and who controls the underlying infrastructure.

Read Full Article

AMIS TAKE

Fortinet’s SASE Summit made one point I keep coming back to: organizations “must therefore secure a growing population of non-human identities as carefully as they secure employees, contractors, and devices.” AI agents are identities, and they need to be governed like one. The summit also warned that “organizations cannot address every new requirement with yet another standalone product.” I’ve watched this happen more and more since 2019. Organizations look at the market before they look at their own tooling. The “best” product changes often, so environments fill up with overlapping tools that each solve one problem, and staff spend more time maintaining them. Real unification means products integrated deeply enough that you can see every communication end to end, human or non-human. Most organizations can’t do that for their AI agents yet.

Read the full AMIS take

HARDWARE / FORTIGATE 1200G

Fortinet Expands FortiGate G Series with 1200G and FortiSASE Outpost to Advance Firewall and SASE Convergence

Fortinet announced the FortiGate 1200G series with FortiSASE Outpost, combining ASIC-powered threat protection with local SASE enforcement. The announcement highlights 10G, 25G, and 100G connectivity and the ability to keep selected inspection and processing in customer-controlled environments.

Read Full Article

AMIS TAKE

The 1200G is for organizations that need high port count at high bandwidth, heavy inspection and encryption for dense offices, or low-latency, privacy-focused protection for remote users through FortiSASE Outpost. If that describes you, and especially if a 1000F or 1100E is in your rack, now is the time to plan. Running a 1000F, 1100E or 1101E? Send us your model, peak traffic and port needs. We’ll map out your refresh and size it on real inspection numbers.

Read the full AMIS take

FROM NEWS TO YOUR NETWORK

Not sure what applies to you? Let’s talk.

Bring the announcement, the question, or the challenge. AMIS can help you put it in context before you plan your next move.

Talk to AMIS

Alvarado Managed IT Services

Fortinet network equipment and managed IT services.