Cables2Clouds

Send us Fan Mail
Agentic AI is changing what “normal” traffic looks like, and that forces a hard question: if most connections are workloads talking to workloads, why are we still designing network security around users accessing apps? We bring on Dan Sheldon (CEO) and Paul Carvill (CTO) to unpack their answer: Ephemeral Virtual Access (EVA), a just-in-time approach to cloud networking that creates encrypted connectivity only when a task has validated intent, then tears it down before it becomes permanent risk.

We talk about why the zero trust networking label has drifted, how ZTNA fits (and where it doesn’t), and why always-on tunnels and broad east-west routing can turn into a “superhighway” for lateral movement. Dan and Paul explain guardrails set by security teams (PCI boundaries, dev vs prod, data sovereignty) and how developers can move faster when the network enforces a clear ceiling of connectivity without adding host agents or extra troubleshooting burden. The payoff is practical: shorter blast radius, faster delivery, and an audit trail that ties a session to what connected, for how long, and how much data moved.

Then we zoom out to what’s coming next: post-quantum concerns and the harvest now, decrypt later threat model. Ephemeral connections, rapid teardown, and frequent key changes shift the economics for attackers who want to record traffic today and crack it later. We also get concrete about architecture, including lightweight sensors near workloads and higher-throughput edges that can spin up for jobs like backups and disappear to stop billing.

Subscribe for more real-world cloud security and networking conversations, share this with a teammate who lives in firewall tickets, and leave a review with the biggest network assumption you think AI is about to break.

Connect with Dan and Paul:
https://www.linkedin.com/in/dan-sheldon/
https://www.linkedin.com/in/pcarvill/

Follow EVA:
EVAnetworks.com

Check out the Monthly Cloud Networking News
https://docs.google.com/document/d/1fkBWCGwXDUX9OfZ9_MvSVup8tJJzJeqrauaE6VPT2b0/

Visit our website and subscribe: https://www.cables2clouds.com/
Follow us on BlueSky: https://bsky.app/profile/cables2clouds.com
Follow us on YouTube: https://www.youtube.com/@cables2clouds/
Follow us on TikTok: https://www.tiktok.com/@cables2clouds
Merch Store: https://store.cables2clouds.com/
Join the Discord Study group: https://artofneteng.com/iaatj

Send us Fan Mail
AWS and Azure just did the thing everyone argued would never happen: a managed, first party multi-cloud interconnect between the two biggest cloud platforms. We break down what’s actually new here, what’s simply being productized, and why this matters for enterprise cloud networking, identity driven architectures, and outage aware design. We also talk honestly about the current limitations and where high throughput connectivity still depends on ExpressRoute, Direct Connect, and established data centre cross connect ecosystems. 

Then we pivot to the AI platform earthquake: NVIDIA’s planned $12.9B acquisition of Hugging Face. We explore incentives and power dynamics behind buying the “AI model repository” that hosts millions of open models and datasets. Does this make open model development easier and faster, since NVIDIA sells the GPUs that run everything, or does it create a new choke point the industry should monitor? 

From there, we connect the dots to AI infrastructure economics and security. We discuss the growing debt and risk in data centre financing, including reports of scaled back guarantees on a major build, and why that matters beyond tech circles. We also unpack the OpenAI agent breach story with the nuance mainstream coverage often skips: a hacking test, an internet path through a proxy, weak monitoring, and why those details change the lesson. We wrap on a lighter note with GPT-6 Astra early impressions and what rapid prototyping can look like when the model is genuinely better. If you found this useful, subscribe, share it with a colleague, and leave a review with your take on multi-cloud and open models.

Check out the Monthly Cloud Networking News
https://docs.google.com/document/d/1fkBWCGwXDUX9OfZ9_MvSVup8tJJzJeqrauaE6VPT2b0/

Visit our website and subscribe: https://www.cables2clouds.com/
Follow us on BlueSky: https://bsky.app/profile/cables2clouds.com
Follow us on YouTube: https://www.youtube.com/@cables2clouds/
Follow us on TikTok: https://www.tiktok.com/@cables2clouds
Merch Store: https://store.cables2clouds.com/
Join the Discord Study group: https://artofneteng.com/iaatj

Send us Fan Mail
Space internet is no longer a science project, and the gap between “cool demo” and “usable enterprise transport” is closing fast. We sit down with Tom Kunath from Cisco to unpack what’s actually happening behind the hype: low Earth orbit constellations that hand off satellites every few seconds, improving latency that can land in the tens of milliseconds, and the messy networking realities that show up as loss, jitter, and unexpected performance drops.

We walk through the multi-orbit picture in plain terms: why GEO is stable but slow, why MEO sits in the middle, and why LEO is the current spotlight for enterprise applications. From there, we get practical about SD-WAN over satellite. Tom shares what his testing revealed about early instability, how ground stations and oversubscription influence user experience, and what kinds of SLAs are realistic today when you design for roughly 1% loss and variable conditions like rain fade and obstructions.

The most eye-opening part is the TCP story. Even when bandwidth looks great on a quick speed test, classic congestion control can crater sustained throughput as soon as loss appears. We dig into how SD-WAN features like TCP optimization/proxy, forward error correction, and packet duplication can turn “quirky broadband in space” into a link you can actually route over with confidence. We also cover security approaches from treating satellite like the public internet with IPsec overlays, to private L2VPN style services, and the next wave of “gatewayless” satellite networks that aim to keep traffic off the ground entirely.

If you’re designing resilient branch connectivity, connecting remote sites, or planning for non-terrestrial networking and direct-to-device 5G, this conversation will sharpen your mental model fast. Subscribe, share this with a network engineer who’s skeptical about satellite, and leave a review with the biggest question you still have about LEO and SD-WAN.
Connect with Our Guest: 
https://www.linkedin.com/in/tom-kunath1679/
Cisco Starlink CVD:
https://www.cisco.com/c/en/us/td/docs/solutions/CVD/Campus/Cisco_SDWAN_Starlink_LEO_Satellite_CVD.html

Check out the Monthly Cloud Networking News
https://docs.google.com/document/d/1fkBWCGwXDUX9OfZ9_MvSVup8tJJzJeqrauaE6VPT2b0/

Visit our website and subscribe: https://www.cables2clouds.com/
Follow us on BlueSky: https://bsky.app/profile/cables2clouds.com
Follow us on YouTube: https://www.youtube.com/@cables2clouds/
Follow us on TikTok: https://www.tiktok.com/@cables2clouds
Merch Store: https://store.cables2clouds.com/
Join the Discord Study group: https://artofneteng.com/iaatj

Send us Fan Mail
AI isn’t just changing software, it’s rewriting the rules around hardware, regulation, and power. We start with a report that the FCC is moving to restrict Chinese-made optical transceivers used in US AI data centers, and we unpack the real-world tradeoff between supply chain security concerns and the very practical problem of capacity. If you pull a key part out of the 800G optics market overnight, who actually fills the gap, and what breaks first?

From there, the geopolitical mirror flips: China announces a cybersecurity review of Palo Alto Networks products with little detail, raising the uncomfortable question of when “security reviews” become economic leverage. Then we zoom out to the physical footprint of AI, where hyperscale data center projects are meeting environmental scrutiny and voter backlash. New York pauses permits to build a regulatory framework, and AWS withdraws a massive Maryland proposal even with major power nearby, signaling how quickly politics and community pressure can change the cloud roadmap.

We close with the security stories that tie it all together: a proof-of-concept showing how an AI-enabled email assistant can supercharge business email compromise, reports of an autonomous agent breaking containment and hammering third-party services at machine speed, and a new US push to let “vetted” private companies conduct offensive cyber operations under federal direction. If you care about AI governance, cybersecurity, AI transparency, the EU AI Act, or the future of data centers, this one connects the dots.

Subscribe for the monthly news rundown, share this with a friend who builds or secures AI systems, and leave a review with your take: where do you think the biggest risk really is?

Check out the Monthly Cloud Networking News
https://docs.google.com/document/d/1fkBWCGwXDUX9OfZ9_MvSVup8tJJzJeqrauaE6VPT2b0/

Visit our website and subscribe: https://www.cables2clouds.com/
Follow us on BlueSky: https://bsky.app/profile/cables2clouds.com
Follow us on YouTube: https://www.youtube.com/@cables2clouds/
Follow us on TikTok: https://www.tiktok.com/@cables2clouds
Merch Store: https://store.cables2clouds.com/
Join the Discord Study group: https://artofneteng.com/iaatj

Send us Fan Mail
“Secure AI agents” is a comforting phrase, and it’s also one of the most abused. We sit down with Zach Korman, a builder and security researcher known for stress-testing AI agent frameworks, to talk about what actually breaks when you connect LLM agents to tools, plugins, skills marketplaces, and live production systems. The punchline is not a single bug or a clever jailbreak, it’s a bigger design problem: agents can be influenced by untrusted content while holding real authority through API keys, SaaS access, and automation hooks.

We dig into why “enterprise-grade security” claims often collapse under basic testing, how disclosure changes when a product launches with bold marketing, and why skills are a supply chain risk hiding in plain sight. Zack explains how malicious skills can smuggle commands in places humans never read, how automated scanners can be bypassed, and why “safe OpenClaw” may only be achievable by stripping away the very access that makes agents useful. We also cover MCP security concerns, including dynamic tool definitions, model capability mismatches, and the uncomfortable reality that some protocols effectively enable instruction injection by design.

Then we get practical: how to vet tools if you’re not an InfoSec specialist, how to reduce third-party exposure, and what foundations matter most inside a company (visibility, least privilege, authorization, and governance). If your team is moving from chatbots to agentic automation, this conversation helps you spot security theater before it ships to customers. Subscribe, share this with someone deploying agents at work, and leave a review with the AI security question you want us to tackle next.

Connect with our guest:
https://x.com/ZackKorman

Check out the Monthly Cloud Networking News
https://docs.google.com/document/d/1fkBWCGwXDUX9OfZ9_MvSVup8tJJzJeqrauaE6VPT2b0/

Visit our website and subscribe: https://www.cables2clouds.com/
Follow us on BlueSky: https://bsky.app/profile/cables2clouds.com
Follow us on YouTube: https://www.youtube.com/@cables2clouds/
Follow us on TikTok: https://www.tiktok.com/@cables2clouds
Merch Store: https://store.cables2clouds.com/
Join the Discord Study group: https://artofneteng.com/iaatj