383 episodes
- Send us Fan Mail
CISSP Domain 4.1 — apply secure design principles in network architectures, built on a joint federal warning about internet-reachable refrigeration controllers.
Start free: https://www.cisspcybertraining.com/free-cissp-self-study (360 practice questions, a Domain 1 diagnostic, and weekly questions by email)
In early November, five US federal agencies put their names on one joint fact sheet. CISA, the FBI, Health and Human Services, Agriculture, and the Department of Transportation. Four sectors named: Healthcare and Public Health, Chemical, Food and Agriculture, and Transportation Systems.
And what they are warning about is the box that keeps the freezer cold at your grocery store
WHAT THE DOCUMENT SAYS
Actors are targeting electronic temperature controllers and commercial refrigeration controllers — the ones in retail grocery stores and cold storage facilities. Three ways in, and the agencies name all three: authentication bypass, arbitrary code injection, and weak configurations.
Once they are in, they can remotely alter settings and force those units into continuous heating and defrost cycles.
The agencies list what that could do. It could affect the service life of the system. It could create a fire hazard. It could cause loss of inventory. And it could indirectly harm consumers, if products are consumed after they were stored at an unsafe temperature.
A freezer, run backwards, is a fire hazard and a public health problem.
THE SECOND EFFECT, AND THE BEST PHRASE IN THE DOCUMENT
There is one more consequence on their list: loss of visibility into current temperatures. Their phrase for it is a denial-of-view condition.
Think about what that means. The actors do not have to change a single setting. They only have to take away your ability to see. And now the temperature log — the compliance record, the thing you would hand an inspector — is reporting a number that is not true, or is not reporting at all.
You cannot manage what you cannot observe, and an attacker who understands that does not need to break anything.
WHAT THE DOCUMENT IS NOT
This is a warning, not a breach report. There is no victim, no company, no country, no count, and no date of an incident, because the document is not about one event. The verb throughout is "could" — the agencies are describing capability, not casualties.
Press coverage of this fact sheet carries figures the government document does not: store counts, countries, temperatures. None of them are asserted in this episode. If you want them, they belong to the outlets that published them.
WHY THESE CONTROLLERS ARE CONNECTED AT ALL
The device is not a toy. The fact sheet lists exactly what it is for: automated and remote monitoring of temperatures, continuous temperature logging, defrost cycle management, and energy optimization. Four jobs, and every one of them is a business function. The logging is a compliance record. The defrost management keeps the coils from icing over. The energy optimization is somebody's utility bill.
Look at the first one again. REMOTE monitoring. That is not laziness. If you are running a lot of stores, you are not sending a person around to read a thermometer. Remote monitoring is the entire reason the thing was bought.
So it got connected. And the question nobody asked at the time was: connected to what?
Now look at the three vectors again. Authentication bypass. Arbitrary code injection. Weak configurations. Not one of them is exotic. They are the three oldest doors in the building, and they work here because this class of device spent twenty years being designed for a network nobody outside the building could reach.
WHY THIS IS A DOMAIN 4.1 EPISODE
Eight mitigations on the agencies' list. The first two are not about the device at all.
Number one: do not expose control system devices — and they call out controller panels specifically — to the internet. If remote access is necessary, protect it with a firewall or an edge device and keep that updated.
Number two: create a dedicated OT virtual local area network, an OT VLAN, to separate those devices from the business network.
Neither of those is a patch and neither one is a product. They are design decisions, and somebody made them, or didn't make them, probably years ago.
THE MANAGER'S TAKEAWAY
Ask who owns that controller.
Here is how it usually goes. Facilities bought it. A vendor installed it. The vendor needed remote access to support it, so somebody in IT opened a path, and it was a Tuesday and it worked. It is not in the asset inventory. It has never been in a vulnerability scan scope. And it has an address somebody can reach from the other side of the world.
Reachable is a design decision. It is not an accident. The only real question is whether the person who made it was in security.
IN THE TRAINING BLOCK
• What 4.1 actually asks: physical, logical and micro-segmentation in ISC2's own wording; edge networks; traffic flows, north-south and east-west. Domain 4 is 13% of the exam.
• Why a VLAN is a configuration rather than a wall.
• North-south and east-west: a boundary is structurally unable to see traffic that never crosses it, and in a flat network nothing else is looking either. That is the argument for micro-segmentation.
• Three questions to ask about any connection: what can reach it, what it can reach, who decided.
• Six shapes of remote access and how each one quietly widens.
• Denial of view, four ways — dark, stale, lying, and the record that inherits all three
• The management plane: in-band reaches a device over the network it serves, out-of-band over a path that does not depend on it — and that only matters on the one day you need it. Plus edge networks, where peering is a commercial agreement that creates a route.
FOUR PRACTICE QUESTIONS, with the traps taken apart. Question two is the one that gets people.
CHAPTERS
0:00 Five agencies, one fact sheet, one freezer
1:45 The whole thing in thirty seconds
3:33 What these controllers actually do
4:51 Three ways in, and what they have in common
5:25 What they can do once they are there
6:00 Denial of view
6:58 The manager's takeaway
8:19 Who owns that controller
9:15 Free Essentials, Accelerator, Pro and the Cohort
11:30 What Domain 4 actually asks
16:34 Three kinds of segmentation
18:19 North-south, east-west
21:44 Every remote path is a design decision
25:13 Denial of view, four ways
28:18 Question 1
31:19 Question 2 — the one that gets people
34:43 Question 3
37:58 Question 4
41:09 The pattern across all four
LINKS
Free CISSP Essentials: https://www.cisspcybertraining.com/free-cissp-self-study
All episodes and the written walkthrough: https://www.cisspcybertraining.com/podcast
CISSP Sprint Cohort: https://www.cisspcybertraining.com/cohort-sprint-pag
SOURCE — the joint fact sheet itself, TLP:CLEAR, 5 November 2026:
https://www.cisa.gov/resources-tools/resources/us-government-partners-urge-hardening-industrial-refrigeration-and-cold-chain-monitoring-systems
Gain exclusive access to 360 FREE CISSP Practice Questions at FreeCISSPQuestions.com and have them delivered directly to your inbox! Don’t miss this valuable opportunity to strengthen your CISSP exam preparation and boost your chances of certification success.
Join now and start your journey toward CISSP mastery today! - Send us Fan Mail
CISSP Domain 3.9 — design site and facility security controls, built on a real data center power failure where every control in the building worked and the building still went dark.
Start free: https://www.cisspcybertraining.com/free-cissp-self-study (360 practice questions, a Domain 1 diagnostic, weekly questions by email, and the slides from this episode)
In October, a cloud provider's London data center lost its utility power. The batteries picked up the load exactly the way they were supposed to, ran for about five minutes, and then ran out. The generators should have taken over in that window. They never got the signal.
Nobody attacked anybody here. There are no bad guys in this one. No ransomware, no intrusion, nothing for incident response to chase. It was a cable fault outside the building and a logic error inside it.
HOW THE BUILDING WAS BUILT
Start with the design, because the design was good. Two completely independent utility feeds coming in from the power company — they call them A and B. Google's own words for that are a 2N utility power architecture.
Now listen to where that label stops. Behind both of those feeds sits one central bank of high-voltage generators, run by a dual-redundant controller, a PLC. Two of the thing coming in. One of the thing that was supposed to save them. And between the utility going away and the generators coming up, a battery system with about five minutes of runtime. That isn't a lot and it isn't supposed to be. It's a bridge.
So when somebody tells you a site is 2N, the only honest follow-up is: 2N at which layer?
Five days before any of this, they ran a quarterly load test on the generators. A real test, under load. And that test found a problem. So they did the right thing — they scheduled the repair. But to do the repair they had to physically open the breaker on feed B and lock it out, which they did properly: keyed interlock, locked out, tagged out.
And that is the moment the building stopped being 2N on the utility side. For several days — Google doesn't publish exactly how many — that site ran on one utility feed. Google's own words for it are a temporary single-feed, 1N, bridged maintenance posture.
Then the cable broke. An underground cable out on the utility network, outside the building entirely, developed a fault. And it was on feed A. The only feed that was live.
WHY THE GENERATORS NEVER STARTED
The PLC decides when to start the generators, and it decides it by sensing voltage. If the voltage goes away, start the generators. Simple enough.
But those voltage sensors were wired upstream of the feed breakers. On the line side. And feed B — the one that was open for repairs — was still energized upstream, all the way up to that open breaker. The power company was still feeding it. It just wasn't connected to the building anymore.
So the PLC looks at its sensor, sees voltage on feed B, concludes that a healthy utility source is available, and never issues a start command.
Google's own root cause is this: a misconfiguration of the PLC logic that did not verify both an undervoltage condition and physical breaker position for both utility feeders. It checked whether there was voltage. It never checked whether that voltage was actually connected to anything.
And here's the part worth a line in your notes. That PLC was dual-redundant. There were two of them. Both of them were wrong in exactly the same way, because they shared a configuration. Duplicating the box does not duplicate the judgment. That's common-mode failure, and it is the reason a redundancy count is not a resilience measurement.
Five minutes later the batteries were done and the data halls went dark.
TWO NUMBERS, KEPT APART
About ninety minutes with the data halls unpowered. Six hours and eight minutes before the services on top of them were whole again. Both figures are in Google's report and they measure different things. According to The Register, the companies caught in that window included Lloyds, HSBC, Royal Mail and Revolut — Google published no customer list and no customer count.
To their credit, Google published the whole thing: the timeline, the wiring, the root cause, and a list of fixes with dates on them. One of those action items is the most honest line in the document. They committed to making PLC code review part of facility commissioning — which tells you it wasn't.
THE MANAGER'S TAKEAWAY
Two things, and neither one is "test your generators."
They DID test. The quarterly load test is what found the fault in the first place. What nobody did was look at the consequence: fixing that fault meant opening a breaker, and opening that breaker meant the building ran on one feed for days. As far as anybody treated it, that was a maintenance ticket. It wasn't a changed risk posture.
And the second one, the one to carry into the exam room: an automatic control is only as good as the thing it can sense. The generator was perfect. The battery was perfect. The logic was asking a question — is there voltage? — at a place where it could not tell the difference between "the utility is healthy" and "the utility is healthy and no longer connected to us."
Redundancy isn't a label on a drawing. It's a property of a system at a particular moment, and that moment includes whatever your facilities team happens to have open for repairs today.
IN THE TRAINING BLOCK
• The power chain end to end, and the fourth thing almost nobody reads — the logic that decides when to start, which lives on a controls drawing rather than a power one.
• UPS runtime: battery in minutes, flywheel in seconds, and why five minutes is a bridge rather than a weakness.
• Redundancy notation — N, N+1, 2N, 2(N+1) — and why N+1 survives one failure but not one failure during maintenance.
• Concurrent maintainability versus fault tolerance: two different promises that candidates collapse into one.
• The three states of a redundant system — designed, degraded, failed — and why failure almost never arrives on its own.
• Every physical control has a sensing input: fire, cooling, water detection, door alarms, access control vestibules.
• The 3.9 recall content, and the crossovers named out loud — change management is 7.9, risk acceptance is 1.9, resilience and fault tolerance is 7.10.
FOUR CISSP PRACTICE QUESTIONS
Set inside the real incident, so they are hard on the reasoning instead of hard on the reading. Common-mode failure; a FIRST question where the trap is the answer you are proudest of spotting; a MOST DIRECTLY question where the trap is the biggest answer on the list; and a control-class question where the trap is the governance gap the episode itself spent ten minutes describing.
CHAPTERS
0:00 Two power feeds, and the generators never started
1:56 The whole thing in thirty seconds
3:36 A spare, or a complete second path
4:44 How the building was built
6:09 The load test, the open breaker, the cable fault
7:40 Why the controller never issued a start command
9:00 Two controllers, one mistake
10:18 The manager's takeaway
12:33 Free Essentials, Accelerator, Pro and the Cohort
15:51 Domain 3.9 and the power chain
19:03 Redundancy, and where the label stops
22:01 The three states of a redundant system
23:37 Every control has a sensing input
29:17 Question 1
32:20 Question 2 — the one that gets people
35:23 Question 3
38:31 Question 4
41:43 The pattern across all four
LINKS
Free CISSP Essentials — 360 practice questions, a Domain 1 diagnostic, weekly questions by email, 8+ hours of rapid review audio and video, and the slides from this episode:
https://www.cisspcybertraining.com/free-cissp-self-study
All episodes and the written walkthrough:
https://www.cisspcybertraining.com/podcast
CISSP Sprint Cohort — for people with a test date on the calendar:
https://www.cisspcybertraining.com/cohort-sprint-page
SOURCES
PRIMARY — Google Cloud's own incident report for europe-west2, Incident Xs2XTHzAVC9YnafnPyzj. Every figure above comes from here unless stated otherwise:
https://status.cloud.google.com/incidents/Xs2XTHzAVC9YnafnPyzj
SECONDARY — The Register, 5 November 2026, Simon Sharwood. The four customer names come from here, NOT from Google:
https://www.theregister.com/off-prem/2026/11/05/when-datacenters-lose-power-generators-should-kick-in-that-didnt-happen-for-google-in-london/5307739
Gain exclusive access to 360 FREE CISSP Practice Questions at FreeCISSPQuestions.com and have them delivered directly to your inbox! Don’t miss this valuable opportunity to strengthen your CISSP exam preparation and boost your chances of certification success.
Join now and start your journey toward CISSP mastery today! - Send us Fan Mail
CISSP Domain 2.6 — data security controls, compliance requirements, and the three states of data, built on an AI hosting exposure that encryption could never have reached, because the data was exposed while it was being used.
Start free: https://www.cisspcybertraining.com/free-cissp-self-study (360 practice questions, a Domain 1 diagnostic, weekly questions by email, and the slides from this episode)
One hundred and fifty-five answers went to the wrong people. Nobody broke in.
In late August, Cloudflare launched the GLM-5.3 model on Workers AI. The model advertised a context window of just over one million tokens. The memory pool behind it — the prefill KV cache — held about six hundred and eighty thousand. The pool was smaller than the thing it was advertised to hold.
Picture a shared workbench with a scratch pad on it. You sit down, you do your working-out, you get up and leave. In a well-run shop somebody wipes that pad before the next person sits down. That pad is memory, and wiping it is the control.
When a request came in too big for the pad, the system marked it rejected, which was correct, and then passed the pad downstream anyway. The next stage read it, found the previous request's working still on it, and handed that back as a successful answer. Not an error. An HTTP 200.
155 responses were affected, every one from a prompt over the pool's limit. Cloudflare said only that a small number of customers were involved and published no customer count. It found no evidence of deliberate exploitation. This was an exposure — nothing was taken, nobody broke in — and the misconfigured value and the serving-engine behavior both sat in the hosting stack, not with the model's maker. The advertised context length itself was accurate.
Security researcher Gadi Evron of Knostic reported it through HackerOne. Cloudflare began verifying it the next morning and reproduced it from two independent accounts within twenty-five minutes of declaring the incident — and the model was offline about fifteen hours after the report landed. Root cause thirty-five minutes later, corrected configuration seventeen minutes after that. It raised the pool to roughly twice the advertised window, moved retirement of rejected requests to the admission queue before any cache transfer, added a deployment check comparing pool size to context length, and published the whole timeline itself.
Now the part that turns it into exam content. Encryption at rest and encryption in transit were never going to touch this one. Not because anything was wrong with them. Because neither one has a single thing to say about data while it is being used.
And they knew. Cloudflare knew before launch that the two numbers disagreed. It assumed the engine would handle it cleanly, and what it validated was the status code and the performance — not the answer coming back. The engine said 200 OK, so it looked fine. It was not a secret. It was an assumption nobody had validated, standing in for a control nobody owned.
A control is not a control because of what it is called. It is a control because of what fails when it is wrong.
In this episode:
• The three states of data — at rest, in transit, in use — and why the third is the one you get tested on. At rest and in transit are cryptography questions. In use is an isolation question, and cryptography alone does not reach it. Confidential computing is the exception.
• The data-in-use control vocabulary: memory isolation, zeroization between requests, object reuse controls, confidential computing, homomorphic encryption and admission control. Object reuse is the classic ISC2 phrasing for this failure shape.
• Scoping and tailoring, and the precision trap. Scoping is not a step before tailoring — it is one of the six tailoring actions NIST SP 800-53B names, and an option that sequences them is the distractor.
• Standards selection and baselines: ISO 27001, NIST SP 800-53 and 53B, CSF 2.0, CIS, the Cloud Controls Matrix, PCI DSS and SOC 2, taught by differentiator rather than contents. A baseline is a floor, not a fit.
• Data protection methods — DRM, DLP, CASB, tokenization and masking — and why not one of them would have caught this. Every one reads a content flow.
• Multi-tenancy and shared responsibility, and the control each side assumed the other already owned.
• The crossover, named out loud: "should we accept this?" is Domain 1. "How is it enforced?" is Domain 3. "What does this data require?" is 2.6, and that is this episode.
• Four CISSP practice questions and the trap inside each one. Question two is the one that gets people.
The line to carry into the exam room: Domain 2.6 does not ask you to list encryption types. It asks which controls this data, in this state, in this environment actually requires — and then whether anybody owns them as requirements.
LINKS
Free CISSP Essentials:
https://www.cisspcybertraining.com/free-cissp-self-study
All episodes and the written walkthrough:
https://www.cisspcybertraining.com/podcast
The CISSP Sprint Cohort — 8 weeks, live, paced to your exam date, with an accountability partner:
https://www.cisspcybertraining.com/cohort-sprint-page
SOURCES
Primary — Cloudflare's own incident writeup, published late October 2026:
"Cross-request KV cache exposure in the GLM-5.3 Workers AI model, and our remediation with SGLang"
https://blog.cloudflare.com/kv-cache-exposure-remediation
Reported by security researcher Gadi Evron, Knostic, via HackerOne.
Shon Gerber, CISSP — former Commander, 177th Information Aggressor Squadron. Former CISO, fractional CISO, founder of CISSP Cyber Training.
Gain exclusive access to 360 FREE CISSP Practice Questions at FreeCISSPQuestions.com and have them delivered directly to your inbox! Don’t miss this valuable opportunity to strengthen your CISSP exam preparation and boost your chances of certification success.
Join now and start your journey toward CISSP mastery today! - Send us Fan Mail
A company found a concealed recording device in the room where its executives meet. They pulled it, imaged it, reviewed the camera footage, went through the employee's internal tool usage, found a second device in a bag he had left in another room, terminated him and referred the matter to the FBI.
Every one of those steps was reasonable. And somewhere in the middle of that list, the investigation stopped being theirs.
This episode is CISSP Domain 1.6 — requirements for investigation types. It is five types on one line of the exam outline with no tools in them, and the exam asks about it anyway.
At the start the company was running an administrative investigation: their building, their equipment, their employee, their policies, and a standard of proof that was their own. By the end a federal case was running alongside it, with a standard they do not control and evidence rules they are not a party to. Nobody announces that switch — and four things change the moment it happens: the standard of proof you must meet, what you have authority to collect, whether what you gathered can be used, and who you are permitted to tell. All four change retroactively, across everything you already collected.
In this episode:
· All five investigation types — administrative, criminal, civil, regulatory, industry standards — and what each one can actually do to you
· Why the type follows whose AUTHORITY the inquiry runs under, not the conduct and not the severity
· Burden of proof by type, and why administrative is the one standard the organization sets for itself
· Chain of custody worked step by step from the real case, including the bag that cannot be un-opened
· Why a forensic image that produced no readable data is still a finding, and why process produces admissible evidence while content produces a conclusion
· The five characteristics of evidence, and why only four of them are properties of the evidence
· Civil preservation: the duty attaches when litigation is reasonably anticipated, not when a complaint is served
· Regulatory: the regulator sets the scope and can compel production, so your own investigation discharges nothing
· Industry standards: a contractual obligation, where the contract names who is permitted to do the work
· What a manager decides before any of it — naming the type out loud, setting the referral trigger in advance, and collecting as though the type will change
It closes with four scenario questions, spread deliberately across all five types. Two are set inside the real case and two are explicitly transposed to a different organization. Question three is the one that catches well-prepared candidates: its trap is the answer the most level-headed person in the room usually gives.
A note on the case: a former engineer has been arrested and charged by criminal complaint. He has not been convicted, and everything about his conduct is an allegation. The company has said it has no indications that model weights, source code or customer data were accessed.
Source: U.S. Department of Justice, Office of Public Affairs press release, October 2026.
Newsroom: https://www.justice.gov/news
Gain exclusive access to 360 FREE CISSP Practice Questions at FreeCISSPQuestions.com and have them delivered directly to your inbox.
Go to CISSPCyberTraining.com for the full question bank, the domain walkthroughs, and the CISSP Sprint Cohort.
Gain exclusive access to 360 FREE CISSP Practice Questions at FreeCISSPQuestions.com and have them delivered directly to your inbox! Don’t miss this valuable opportunity to strengthen your CISSP exam preparation and boost your chances of certification success.
Join now and start your journey toward CISSP mastery today! - Send us Fan Mail
A “boring” GitHub Actions workflow file should never be able to empty your secrets vault, yet that is exactly what the Ghost Action campaign shows. We walk through GitGuardian’s research and the real mechanics of how attackers use stolen but valid developer tokens to add or modify workflow YAML, scrape secret names from your own repo, and exfiltrate CI/CD secrets to an external endpoint. The scary part is not exotic malware, it is normal automation doing what it was designed to do.
From a CISSP Domain 8.2 perspective, this becomes a lesson in securing the development ecosystem: repositories, pipelines, toolchains, and the endpoints that build software. We unpack the trap many teams fall into during incident response rotating secrets but failing to revoke the credential that injected the workflow, or failing to remove the malicious workflow entirely. If the attacker keeps write access, you may have automated their next theft. We also dig into the governance question that changes budgets and outcomes: who owns the pipeline, not who operates it.
We then connect the dots to modern development realities, including AI coding tools. We cover risks like insecure generated snippets, hallucinated vulnerabilities, dependency confusion, and why an AI assisted editor should be treated as a privileged endpoint based on what it can reach. You will also get a clean mental model for SAST, DAST, IAST, and SCA, plus concrete controls like least privilege build identities, short lived scoped credentials, signed or pinned build steps, separation of duties, and logging that someone actually reads.
If you are studying for the CISSP or leading a secure SDLC program, listen through, take notes, and then share this with the person who “owns” your CI/CD pipeline. Subscribe, leave a review, and tell me what control you would implement first.
Gain exclusive access to 360 FREE CISSP Practice Questions at FreeCISSPQuestions.com and have them delivered directly to your inbox! Don’t miss this valuable opportunity to strengthen your CISSP exam preparation and boost your chances of certification success.
Join now and start your journey toward CISSP mastery today!
More Courses podcasts
Trending Courses podcasts
About CISSP Cyber Training Podcast - CISSP Training Program
Join Shon Gerber on his weekly CISSP Cyber Training podcast, where his extensive 23-year background in cybersecurity shines through. With a rich history spanning corporate sectors, government roles, and academic positions, Shon imparts the essential insights and advice necessary to conquer the CISSP exam. His expertise is not just theoretical; as a CISSP credential holder since 2009, Shon translates his deep understanding into actionable training. Each episode is packed with invaluable security strategies and tips that you can implement right away, giving you an edge in the cybersecurity realm. Tune in and take the reins of your cybersecurity journey—let’s ride into excellence together! 🚀
Podcast websiteListen to CISSP Cyber Training Podcast - CISSP Training Program, Deep Dive: CFA® Level I Prep 2026 and many other podcasts from around the world with the radio.net app

Get the free radio.net app
- Stations and podcasts to bookmark
- Stream via Wi-Fi or Bluetooth
- Supports Carplay & Android Auto
- Many other app features











