What if the most devastating move in patent law isn’t a bold block — but a quiet allowance?
The U.S. Court of Appeals for the Federal Circuit just affirmed a PTAB smackdown on Centripetal Networks’ U.S. Patent 10,284,526, declaring its core claims unpatentable as anticipated. We’re talking network security here: decrypting packets, peeking inside, acting accordingly, re-encrypting, and shipping them off. Keysight Technologies challenged it via IPR, and the PTAB — backed now by Judge Lourie and crew — said prior art like Cisco’s IronPort AsyncOS 7.1 User Guide (IPUG) beats it hands down.
But here’s the rub. Centripetal screamed that the PTAB mangled claim construction. The disputed bit? “Performing a corresponding action” on each decrypted packet. PTAB said nah, that includes just letting the packet through — no drama needed. Centripetal? They called it superfluous, arguing it’d gut the later “transmitting” step.
Why Does ‘Allowing’ a Packet Suddenly Count as an Action?
Look. Patents live or die on words. Claim 1 demands decrypt, act, re-encrypt, transmit. Centripetal insisted “action” meant something muscular — like blocking threats — not passive permission.
The PTAB — and now CAFC — saw it differently. Straight from the spec: blocking’s an example, so its flip side, allowing, fits too. Judge Lourie nailed it:
“Allowing a packet is the opposite of blocking it and would naturally be considered an action.” – CAFC
Boom. No wiggle room. And that transmitting step? It’s downstream — re-encrypt first (post-allow), then send. Distinct steps, no overlap. Centripetal’s surplus argument? Shot down. Substantial evidence from IPUG shows policies that block, redirect, or — crucially — allow, followed by encryption and dispatch.
IPUG spells it out: apply access rules to decrypted traffic. If allowed, encrypt and send to client. That’s anticipation, full stop.
This isn’t just semantics. It’s architectural. Network security patents often hinge on these flowcharts of inspect-act-forward. But if “act” balloons to include do-nothing-passthrough, prior art — user manuals, even — swallows them whole.
Was Cisco’s User Guide Really ‘Public’ Enough to Kill the Patent?
Centripetal didn’t stop at construction. They hit the prior art status hard. IPUG? A webpage. Not Google-indexed, they claimed — no real public access under 35 U.S.C. §102.
CAFC shrugged. Substantial evidence: Cisco hosted it openly, no logins, downloadable. Keysight’s expert testified to Wayback Machine captures pre-Centripetal’s priority date. Searchable via Cisco’s site? Check. Industry pros could’ve grabbed it? Yup.
Public accessibility doesn’t demand Google stardom — just reasonable chance of discovery by interested eyes. Precedents like Blue Calypso back this: niche sites count if targeted searches hit ‘em.
Centripetal’s gripe feels like grasping. In cybersecurity, where docs like IPUG float freely among engineers, this was primed for takedown.
And the other grounds? Obviousness over Altman, combos? Moot, since anticipation alone dooms claim 1 — and it’s representative.
The Hidden Shift: Cybersecurity Patents on Shaky Ground
Dig deeper. This echoes the post-Alice quake in software patents. Functional claims — decrypt, analyze, act vaguely — invite invalidity. Centripetal’s spec lists blocks, tags, drops, but doesn’t cabin “action.” Courts hate that ambiguity now.
My take? Unique angle the rulings skip: this presages a cull in network patents. Think firewalls, IDS/IPS — all decrypt-inspect-forward. User guides from Cisco, IronPort era? Goldmines of prior art. Expect IPR filings to spike, PTAB wielding these like a hammer.
Bold prediction: by 2026, 30% more cybersecurity assertions invalidated on anticipation. Why? Guides aren’t hiding; they’re online relics. Patent drafters, take note — specify actions ruthlessly, or watch packets (and claims) slip away.
Corporate spin from Centripetal? Silent so far, but expect appeals or reissues. Keysight? Quiet win, bolstering defenses.
Here’s the thing. PTAB’s workload balloons with tech IPRs — over 1,000 yearly now. CAFC affirmances like this signal stability: broad constructions, easy anticipation finds. No mercy for loose language.
Wander a bit: remember Enfish? Self-referential databases saved by specifics. Centripetal lacked that. Lesson etched in silicon.
Why Should Network Engineers Care About This Patent Dust-Up?
You’re building next-gen firewalls. This ruling? Your prior art shield thickens. Challengers cite guides — your own docs might haunt you.
Drafters: “action” begs scrutiny. Narrow it. Or risk PTAB’s gaze.
Investors in cyber IP? Bearish short-term. Validity rates dip.
🧬 Related Insights
- Read more: Disney’s Mandalorian Moment: Forcing AI Pirates to Pay the Piper
- Read more: EU’s AI Act Bombshell: 10²³ FLOPs Threshold Redefines Everything
Frequently Asked Questions
What does the Centripetal v. Keysight CAFC decision mean?
It upholds PTAB invalidation of Centripetal’s network security patent claims as anticipated by prior art, mainly over broad claim construction of “performing a corresponding action.”
Is Cisco IPUG considered prior art after this ruling?
Yes — publicly accessible via Cisco’s site pre-priority date, per substantial evidence including Wayback captures and expert testimony.
How does this affect cybersecurity patent strategies?
Expect tighter claim language; vague functional steps like generic “actions” vulnerable to user guides as anticipating references.
