Post-quantum cryptography for federal networks: securing Cisco from boot to transport

Key takeaways

• The threat is harvest now, decrypt later: encrypted federal traffic captured today can be unlocked the moment a quantum computer arrives, so any data with a multi-decade secrecy requirement is already at risk.
• Quantum-safe is a full-stack property, not a single setting. It touches the boot chain, the identity and certificate layer, and every encrypted transport tunnel between sites, clouds, and remote users.
• NSA's CNSA 2.0 sets binding timelines for national security systems, and NIST has finalized the underlying algorithms (ML-KEM, ML-DSA, SLH-DSA), so the standards uncertainty that justified waiting is gone.
• Most of the transition is software and configuration on platforms you already own, with hybrid key exchange as the practical first step. Hardware replacement should ride your normal refresh cycle, not an emergency rip-and-replace.
• Inventory first. Map cryptography to data sensitivity, prioritize transport over contested networks, and confirm TAA status part number by part number, plus whether a platform completed DoDIN APL certification where a DoD program still asks for it, on every replacement SKU before it goes on the order.

The clock that started years ago

Quantum computers cannot break today's encryption. They do not need to be able to yet, because the attack that matters does not require a working quantum machine at all. It is called harvest now, decrypt later: an adversary captures encrypted federal traffic as it crosses a backbone or a public network today, stores it cheaply, and waits. The day a cryptographically relevant quantum computer comes online, every session in that archive that relied on classic public-key cryptography becomes readable. For a SaaS password that rotates monthly, that is a non-event. For intelligence, weapons design, personnel records, or treaty data that must stay secret for thirty years, the breach has effectively already happened.

That single fact is what reframes the whole conversation for government buyers. The relevant question is not when a quantum computer arrives. It is how long your data has to stay confidential, and whether the links carrying it today are already feeding someone's archive. The NSA has been explicit that the migration window for national security systems is now, and the foundational standards are no longer in flux. In August 2024 NIST published the first finalized post-quantum standards, ML-KEM for key establishment and ML-DSA and SLH-DSA for signatures, issued as FIPS 203, 204, and 205. The excuse that the math might change has expired.

Cisco is one of the largest installed bases in the federal estate, which means the practical migration is, for most agencies, a Cisco migration. The good news is that this is a sequencing problem more than a science problem. The bad news is that agencies that treat it as a someday item will discover the longest-lived secrets were the ones they left exposed the longest.

What 'full-stack' post-quantum actually means

Quantum-safe is not a checkbox you tick on a firewall. It is a property of every layer that touches a key, a certificate, or a session, and a single classical link anywhere in the chain undermines everything upstream of it. On a Cisco network that chain runs from the silicon a device trusts at power-on all the way out to the tunnels carrying traffic between buildings and clouds. If you secure transport but leave image signing on a classical algorithm, you have moved the weak point, not removed it.

It helps to break the stack into three tiers, because each one has a different migration mechanism and a different timeline. Boot and identity is largely a hardware and firmware question. The control plane is a certificate and key-management question. Transport is mostly a configuration and software-train question, which is why it is usually where agencies see the fastest, most visible progress. Sequencing the work means knowing which tier each platform sits in and what it will take to move it.

The three layers, in plain terms:

• Boot and identity: a hardware root of trust and image signing that verify a device only runs Cisco-signed software. This is where platforms like the Cisco Trust Anchor module live, and where quantum-resistant signatures eventually have to replace classical ones so a forged image cannot be slipped onto a router or switch.
• Control plane: management, telemetry, certificate authorities, and the PKI that authenticates devices and operators. This layer negotiates the key exchange everything else depends on, so it has to become PQC-aware before transport can be fully trusted end to end.
• Transport: IPsec site-to-site tunnels and TLS for management and applications adopting hybrid or post-quantum key exchange, plus MACsec on the wire, which already runs on AES-256 symmetric keys and mainly needs the way those keys are distributed kept quantum-safe, so that captured traffic stays protected even against a future quantum attacker.

Boot to transport: where Cisco platforms fit

Start at power-on. A router or switch that cannot prove it is running authentic Cisco software is a quantum problem waiting to compound, because any encryption it performs is only as trustworthy as the firmware underneath it. Cisco's hardware root of trust and secure boot establish that chain today using classical signatures, and the migration path moves those signatures to quantum-resistant algorithms over successive software and hardware generations. For most agencies this layer rides the normal hardware refresh rather than a standalone project, which is the right instinct as long as the platforms carrying your longest-lived data are near the front of the queue. Our networking and switching practices scope exactly that ordering against your installed base.

The transport layer is where the harvest now, decrypt later threat is most directly answered, and where Cisco has moved fastest. On the Catalyst 8000 Series edge routers running IOS XE 26.1 or later, IKEv2 supports hybrid key exchange, combining a classical elliptic-curve exchange with ML-KEM so the session stays secure even if either method is later broken, and an optional mode lets an upgraded hub fall back to classical exchange for spokes that have not caught up. That hybrid approach is deliberate and it is the consensus interim posture across the industry, because it preserves classical security guarantees while adding quantum resistance, and it degrades gracefully. You can roll it across SD-WAN tunnels and data center interconnects without betting the network on a brand-new algorithm alone. Cisco's enforcement platforms, including the Secure Firewall 3100 Series, are the natural place to terminate and inspect that transport at the edge of an enclave, with one scheduling caveat: Cisco's published Secure Firewall roadmap puts hybrid ML-KEM key exchange in FTD 10.5 and ASA 9.25 in late 2026, so on the firewall side this is a software train to plan for rather than a feature you can switch on today.

Between boot and transport sits identity. Device and user authentication, posture, and segmentation policy all flow through the certificate and key infrastructure that platforms like Cisco Identity Services Engine and the broader security stack manage. As that PKI migrates to post-quantum signatures, the certificates issued to every endpoint and tunnel become quantum-resistant too. None of these three layers moves in isolation, which is the entire argument for planning the migration as one program rather than three disconnected upgrades.

Why federal moves first, and why it is a mandate

The commercial world will migrate to post-quantum cryptography eventually, driven by browsers, payment systems, and compliance frameworks. Federal and defense networks do not have the luxury of waiting for that wave, because the data they carry has a confidentiality requirement measured in decades and an adversary set that is actively resourced to collect against it. That asymmetry is exactly why the government is the leading edge of the transition rather than a follower.

The forcing function is CNSA 2.0, the NSA's Commercial National Security Algorithm Suite 2.0. NSA publishes CNSA 2.0 as an advisory, but CNSSP 15 now incorporates it and makes it the requirement for national security systems, with new NSS deployments expected to comply from January 1, 2027. Its milestones are dated: traditional networking equipment such as routers and VPN gateways should support and prefer CNSA 2.0 algorithms by 2026 and use them exclusively by 2030, earlier than most other categories precisely because it is the connective tissue everything else rides on. Vendors that sell into the national security space are expected to deliver against those dates, which is why Cisco's roadmap is aligned to them rather than to a vague future. Agencies that wrap this requirement into their existing controls catalog will find it dovetails with the cryptographic protection families in frameworks already in force, and DoD environments will see it surface in the hardening guidance published through the DoD Cyber Exchange STIGs.

There is also a procurement dimension that is easy to underestimate. Quantum-safe capability becomes a line item in solicitations, an attribute a contracting officer has to be able to verify against a specific SKU and software train, and a thing that has to be defensible in an audit. That turns a cryptography decision into an acquisition decision, and acquisition decisions in government move on their own timelines. Starting the inventory now is partly about the math and partly about leaving room for the paperwork.

Planning the migration: inventory before algorithms

The single most useful move an agency can make this year has nothing to do with cryptography. It is building an honest inventory of where keys live and what data they protect. You cannot prioritize what you have not catalogued, and migrations tend to stall when the team starts swapping algorithms before it knows which links carry the secrets that actually matter. Map your cryptography to data sensitivity first, then let that map drive the order of operations.

With that inventory in hand, the sequencing tends to write itself. Protect the longest-lived secrets first, because they are the ones already accumulating in someone's archive. Prioritize transport that crosses untrusted or contested networks over links that never leave a controlled facility. Sort platforms into three buckets: those that already support quantum-resistant signing and key exchange in their current software, those that need a software train you can schedule, and those that have to be replaced on their normal refresh cycle. The aim is to fold post-quantum capability into hardware you were going to buy anyway, not to manufacture an emergency.

A disciplined plan looks like this:

• Map cryptography to data sensitivity and protect the longest-lived secrets first, since those are the ones harvest now, decrypt later targets.
• Prioritize transport links that cross untrusted, public, or contested networks, where captured traffic is most likely to already be in an archive.
• Enable hybrid key exchange now where the platform supports it, so you gain quantum resistance without abandoning proven classical security.
• Schedule software trains for platforms that need them, and slot hardware replacement into the existing refresh cycle rather than a separate forklift.
• Confirm TAA country of origin per part number and lot as you go, and for DoD networks note which platforms completed DoDIN APL certification before that program sunset on September 30, 2025, with the documentation a contracting officer can verify.

Folding PQC into a normal Cisco lifecycle

The phrase that should set off alarms in any agency is rip and replace. A panicked, all-at-once hardware swap is expensive, disruptive, and almost never necessary, because the bulk of the post-quantum transition happens in software and key-exchange configuration on platforms you already own. The platforms that genuinely need replacing should be retired when they would have been retired anyway, with quantum-safe successors slotted in on the same cadence. That is how you get to compliance without blowing a capital budget or triggering a year of change-control churn.

Lifecycle discipline is what makes this affordable. Knowing each platform's end-of-sale and end-of-support dates lets you align the post-quantum upgrade with hardware that is already aging out, and Cisco's End-of-Life and End-of-Sale policy fixes the milestones so the timing is never a guess. Keeping production gear under active coverage through Smart Net Total Care ensures you actually receive the software trains that carry the quantum-safe features, since an out-of-contract device is one that quietly misses the upgrade that mattered. Our licensing and lifecycle and managed operations teams keep that coverage and those upgrade windows aligned.

This is ultimately a program, not a product purchase, and programs benefit from a partner who has done the sequencing before. Uniqcli scopes the quantum-safe refresh as part of a normal Cisco lifecycle: we identify the platforms that carry your longest-lived data, sequence the software and hardware moves so transport and identity stay in step, and source TAA-compliant options for every replacement, favoring platforms that completed DoDIN APL certification where a DoD program still asks for it, with the country-of-origin documentation your contracting officer needs. When you are ready to put structure and dates around it, our defense practice and government industry teams can scope the migration and return a quote.

Procurement realities the migration runs into

Even a perfectly sequenced technical plan collides with how government actually buys. Post-quantum capability has to be specified, sourced, and verified through whatever purchase path the agency uses, and that adds months that engineers rarely budget for. The platforms you want may carry the quantum-safe software train but still need their TAA status confirmed part number by part number, and for defense networks the cybersecurity checks that DoDIN APL certification used to cover now run through DISA's vendor STIG program, since the APL sunset on September 30, 2025 and its repository is kept only through FY2026. Those are not afterthoughts. They are gates, and they belong in the schedule from day one.

Choosing the purchase path early is what keeps the timeline honest. Many agencies acquire Cisco infrastructure through governmentwide vehicles such as NASA SEWP or the schedules administered by the General Services Administration, and if that is your agency's route, plan the refresh around its ordering cycle. Uniqcli does not hold a contract vehicle today (our GSA MAS application is in progress), so we sell through the paths that do not need one: Government Purchase Card orders under the $15,000 micro-purchase threshold, simplified acquisition under FAR Part 13 up to the $350,000 threshold, FAR-based purchase orders, and WAWF invoicing through PIEE for DoD. The cleanest migrations treat the purchase path as part of the design, not a billing detail discovered at the end. Our procurement practice exists precisely to keep the BOM, the compliance evidence, and the purchase path moving together.

The throughline across all of it is the same one the NSA, NIST, and every serious vendor keep repeating: start now. Not because a quantum computer is imminent, but because the inventory, the sequencing, the software trains, the hardware refresh, and the paperwork all take time you do not get back, and because the traffic crossing your network today is the traffic most likely to be sitting in an adversary's archive when the math finally turns against you.

Cisco products involved

• Cisco Secure Firewall 3100 Series
• Cisco Catalyst 9000 switching
• Cisco SD-WAN
• Cisco Identity Services Engine
• Cisco Trust Anchor / hardware root of trust
• Cisco Smart Net Total Care
• Cisco Nexus data center fabric

Bottom line: Post-quantum cryptography is a procurement and sequencing problem before it is a math problem, and the agencies that inventory and stage the work now are the ones that avoid the forklift later. Talk to Uniqcli about scoping your quantum-safe Cisco refresh.