RiskControl for SONiC: Mitigate now. Upgrade when ready.

When a new vulnerability is found, network teams often face a difficult trade-off: patch quickly and risk disrupting production, or wait for a maintenance window while the network remains exposed.
Fortunately, customers using SONiC as their network OS have a powerful option to control risk. Tetragon agents can monitor linux kernel processes and apply compensating controls directly via eBPF at runtime. Aviz has validated the capability with its Certified Community SONiC (CCS) as well as other commercially available SONiC distros.
Going forward, Aviz Networks will embed Tetragon agents with its regular CCS releases and provide users the control via Aviz ONES to manage the agent lifecycle and any compensating controls that need to be applied across the switch fleet.
Days versus months
Like any software, Network operating systems and the underlying Linux are impacted by CVEs. What has changed is the discovery clock speed and the time available to stop exploitation.
AI-assisted vulnerability research is compressing the time between disclosure and working exploit from months to days, sometimes hours. Network infrastructure is an especially valuable target. Switches and routers sit at critical points in the fabric, with privileged access to traffic, and services.
On the defender side, the clock has not moved. Upgrading production networks means validation, change windows, rollback plans, and coordination across teams and sites. Even well-run organizations measure that in weeks or months.
That leaves an exposure window: attackers move in days, and you patch in months. You can’t close it by patching faster alone, and you shouldn’t rush unvalidated upgrades into critical infrastructure just to beat the clock.
The industry is converging on the same answer. Open source Tetragon, leveraging eBPF, applies targeted runtime controls to a running switch, decoupling compliance and risk reduction, from the upgrade cycle.
At Aviz, we are bringing that model to every SONiC environment, on any hardware, today.
A three step approach
1. Assess in minutes, not days
Before you can control a risk, you have to understand it. When a CVE lands, the first questions are always the same. Does it affect my network? Which devices, which releases, which configurations? Is the vulnerable code path reachable in our deployment? What is the blast radius? Researching that manually can take days and weeks..
Aviz Network Copilot is an agentic platform customers can use to build their own assessment workflow. It correlates the CVE against your actual inventory, software versions, and configuration, and returns actionable insight within minutes:
- Whether you are exposed, and how seriously
- Which devices and services sit in the blast radius
- The available mitigation path and suggested maintenance windows for devices
Your team spends its time deciding, not digging.
2. Control risk immediately with Tetragon
Knowing your exposure is half the job. The other half is minimizing it now, without waiting for the upgrade.
Tetragon is an open source eBPF-based security and observability engine. It leverages the in-built capabilities of the Linux kernel, so it can see and enforce policy at the exact point where a process opens a file, makes a syscall, or executes a binary. Several properties matter for network infrastructure:
- No reboot, no traffic disruption. Policies load dynamically into a running kernel.
- Targeted actions. A control can be as narrow as “this process must never touch this file” or “this daemon must never spawn a shell.” Narrow rules mean very low false positive rates.
- Low overhead. eBPF programs are verified before they run and execute as efficient native code, so the control plane stays healthy.
- Works across SONiC. SONiC runs on the Linux kernel, and Tetragon observes at the kernel level, so it covers the host and every SONiC container, whether it’s Aviz Certified Community SONiC, or another distribution.
This is why Tetragon will be enabled by default in the upcoming Aviz CCS 2.0 release, and why it can also be applied to any other SONiC environment you run.
Tetragon is not a replacement for patching. It is a mitigating policy that puts risk control squarely in the network operator's hands. Don’t wait for others to protect the life of your network.
An example of what a Tetragon policy looks like: ‘pedit COW’ CVE-2026-46331
CVE-2026-46331 is a Linux kernel vulnerability in the act_pedit traffic-control module. Under the right conditions, an attacker can exploit a memory-handling flaw to corrupt cached data, potentially gaining elevated privileges or crashing the system. Rather than leave that exposure unaddressed while waiting for a validated upgrade, Tetragon can block attempts to load the module, preventing the vulnerable code from being activated.
JSON# CVE mitigation policy: CVE-2026-46331 "pedit COW" (net/sched act_pedit). # # Applies to SONiC 202205 through 202605/master (kernels 5.10.179, 5.10.218, # 6.1.123, 6.12.41; fixed upstream in 5.10.260 / 6.1.177 / 6.12.94). # # The bug: act_pedit computes the copy-on-write range once, before its key # loop, so a pedit rule with typed (header-relative) keys can write past the # COW'd region into shared page-cache pages, e.g. a setuid binary's pages. # Exploitation needs CAP_NET_ADMIN (normally obtained by an unprivileged user # through a new user namespace) to install a tc pedit action. # apiVersion: cilium.io/v1alpha1 kind: TracingPolicy metadata: name: "cve-2026-46331-pedit" spec: lsmhooks: - hook: "kernel_module_request" args: - index: 0 type: "string" selectors: - matchArgs: - index: 0 operator: "Equal" values: - "act_pedit" # 5.10 / 6.1 kernels (202205-202505) - "net-act-pedit" # 6.12 kernels (202511, 202605, master) matchActions: - action: "Override" # hook returns -EPERM: modprobe never runs argError: -1 - action: "Sigkill"
The vulnerability still exists in the code, but the exploits attack vector is gone. The services keep running and the switch keeps forwarding, providing confidence that the network is secure, and ensuring compliance is maintained.
Controls you can trust
Runtime controls on production infrastructure have to be safe to adopt. That’s why we recommend a staged workflow:
- Monitor. Deploy the policy in observe-only mode. Confirm it sees what you expect and nothing legitimate trips it.
- Enforce. Once validated, switch on blocking.
- Retire. After the permanent patch or upgrade is applied, remove the control. It was a bridge, and it shouldn’t linger as undocumented drift.
At every step the control is visible, reversible, and tied to the CVE it addresses.
Managing it all with Aviz ONES
Controls only help if you can operate them across a real fleet. Aviz ONES gives you one place to manage both your SONiC distributions and your Tetragon mitigation policies: deploy, monitor, enforce, roll back, and retire, across devices and sites, including multivendor environments.
Be in control of your risk
A new CVE no longer has to force a choice between exposure and a rushed upgrade.
- Assess the impact in minutes with Aviz NCP.
- Mitigate immediately with Tetragon controls in the kernel, deployed via ONES.
- Upgrade with ONES when the new release is validated and you are ready.
You get everything that makes SONiC compelling: open source, multivendor choice, and flexibility. You also keep the enterprise support, lifecycle, security, and operations experience your business depends on.
Security without the rush
Want to explore how runtime CVE mitigation could fit into your SONiC environment? Talk to our team here at Aviz. We’d welcome a technical conversation about your requirements, and how RiskControl can help.
Frequently Asked Questions
What is RiskControl for SONiC?
RiskControl is a three step approach from Aviz Networks for handling new vulnerabilities in SONiC. You assess the impact in minutes with Aviz NCP, mitigate immediately with Tetragon controls in the kernel deployed via ONES, and upgrade with ONES when the new release is validated and you are ready.
How can network teams reduce CVE risk on SONiC switches without rushing an upgrade?
Customers using SONiC as their network OS can use Tetragon agents to monitor Linux kernel processes and apply compensating controls directly via eBPF at runtime. This decouples compliance and risk reduction from the upgrade cycle.
What is Tetragon?
Tetragon is an open source eBPF-based security and observability engine. It leverages the in-built capabilities of the Linux kernel, so it can see and enforce policy at the exact point where a process opens a file, makes a syscall, or executes a binary.
Does applying a Tetragon policy require a reboot or disrupt traffic?
No. Policies load dynamically into a running kernel, with no reboot and no traffic disruption.
Does Tetragon work on SONiC distributions other than Aviz CCS?
Yes. SONiC runs on the Linux kernel, and Tetragon observes at the kernel level, so it covers the host and every SONiC container, whether it is Aviz Certified Community SONiC or another distribution. Tetragon will also be enabled by default in the upcoming Aviz CCS 2.0 release.
Is Tetragon a replacement for patching?
No. Tetragon is not a replacement for patching. It is a mitigating policy that puts risk control squarely in the network operator’s hands.
How does Aviz Network Copilot help assess CVE impact?
Aviz Network Copilot correlates the CVE against your actual inventory, software versions, and configuration, and returns actionable insight within minutes: whether you are exposed and how seriously, which devices and services sit in the blast radius, and the available mitigation path with suggested maintenance windows.
How does Aviz ONES help manage Tetragon policies across a fleet?
Aviz ONES gives you one place to manage both your SONiC distributions and your Tetragon mitigation policies: deploy, monitor, enforce, roll back, and retire, across devices and sites, including multi-vendor environments.
Mitigate SONiC vulnerabilities without disrupting network operations or rushing upgrades. Explore how Aviz Networks helps you reduce risk with runtime security controls.





