Security has been a board-level concern for a while, but the question is shifting from ‘how secure are we?’ to ‘can we keep ahead of the threats?’. Staying a step ahead means reacting faster, and that depends on a simpler estate.
AI-assisted research is shrinking the time between a vulnerability being found and being exploited, and defenders are racing to keep up. That race is hardest for teams managing network and security equipment from several vendors, each with its own commands and patch cycle. One vulnerability can mean several change windows, and every one of them is time an attacker can use.
Speed depends on simplicity. The business asks haven’t changed (lower cost, less complexity, tighter control), but the pressure behind them has. Software-defined networking with NSX on VMware Cloud Foundation (VCF) delivers all three, and it does so on hosts you already own.
Cost: a smaller physical network bill
You will see the saving at the next networking refresh. Every host a business has bought already carries CPU and memory, and NSX uses a share of it to run switching, routing and, with a vDefend security licence, firewalling and threat prevention. Traffic between workloads no longer has to pass through dedicated physical routers and firewalls, so the physical network can be simpler, with fewer premium features to buy. That changes how much you need and what the refresh could cost.
Timing matters. A hardware refresh or a networking renewal two years out is the time to start looking at adopting software-defined networking. Switching and routing already come with VCF, and firewalling and threat prevention come with the one vDefend licence added on top, so the question at renewal becomes how much physical capability the environment still needs. You do not have to make every decision up front. VCF works with your existing physical network, and because NSX comes bundled with VCF, you can plan to adopt that at your own pace in line with your renewals.
Complexity: one policy, one patch cycle
Complexity adds time and increases risk. On a physical network, engineers configure each device separately, so consistency depends on people entering the same commands every time, and every new rack adds more devices to look after. Estates that change often feel this most. Static estates feel it at patch time, because each vendor updates on its own schedule, and a mixed estate runs several schedules at once.
Software-defined networking replaces that with policy. A team defines the network it needs once, and the policy applies the same way everywhere it is scoped: one cluster, several, or across sites. Patching follows the same logic. VCF updates the networking layer alongside the rest of the platform through a single upgrade workflow. From VCF 9.1, teams can also apply more security fixes to hosts and vCenter with little or no disruption. Your infrastructure team can run much of this from vSphere, the tool it already knows, which shortens the learning curve.
When networking becomes a software attribute of the platform, the silos between networking and infrastructure teams dissolve into a single platform team. I made the wider case for that in a previous article, and the same push towards standardisation runs through VCF deployments more generally.
Control: protection that follows the workload
The policy belongs to security; the mechanism belongs to the existing platform. Perimeter firewalls remain essential and nothing here replaces them, but they only protect traffic that crosses a network boundary. Traffic between two workloads on the same network never reaches them. On a flat network with one hundred virtual machines, that means one compromised workload, such as a virtual desktop where someone clicks the wrong link, has a clear path to the other ninety-nine.
vDefend closes that gap by enforcing policy at each workload. A rule states which machines may talk to which, and it applies whether they share a network or not. There is no security agent to install and patch on each workload, and the rule follows the virtual machine when it moves. The firewall switches on with a licence inside NSX, so there is no separate product to deploy. If a workload is compromised, a quarantine policy isolates it with a single tag, and its neighbours carry on.
Segmentation works as a ladder. Customers start with shared services, then separate environments such as production and development, then applications, and only then the tiers inside an application. Out of the box everything is allowed, so your team adds protection as it publishes policy, stage by stage. Where an organisation starts will depend on its sector and specific requirements.
The vDefend Security Services Platform, deployed alongside NSX, maps which workloads talk to which and recommends policy, making the first rung a lot less daunting. Broadcom’s guided workflows take your team through the stages one at a time: Distributed Firewall (DFW) 1-2-3-4 for segmentation and Advanced Threat Prevention (ATP) 1-2-3 for threat prevention. Each step adds protection and gives you progress to report to the board.
VCF 9 also adds Virtual Private Clouds (VPCs), a model familiar to public cloud users. A VPC gives an application or tenant its own bounded space of private networks. VCF provides the isolation, and vDefend adds firewalling inside and around it. VCF 9.1 adds ready-made security profiles built on vDefend, delivered through VCF Automation, so application teams get protection within guardrails the infrastructure team sets.
Speed: buying time while the real patch lands
Patching still matters, and so does buying time. vDefend’s Advanced Threat Prevention capability includes virtual patching: protection that blocks known exploit attempts against vulnerable workloads until your team applies the real patch. Tag a workload as vulnerable and the protection applies automatically; remove the tag after patching and it withdraws itself. Services that are hard to take offline, such as critical national infrastructure, healthcare or supply chains, gain the most from that extra time. For web-facing applications, Avi Load Balancer’s web application firewall adds the same protection at the application layer.
Virtual patching also changes the conversation between infrastructure and application teams. A fix for a critical flaw cannot always go straight to production: applications need testing, vendors have to ship their own updates, and some software never receives an official fix.
The PHP2Shell flaws in October 2026 showed all three. Amazon recorded an exploit attempt four days before the fix. Vendors embed PHP in products that customers cannot update themselves, many older PHP versions are no longer supported, and PHP published a further advisory within two days.
Without a third option, teams either release untested changes or take services offline, as many organisations did that week. Virtual patching blocks the known exploit as soon as protection exists, so application teams can test and deploy the real fix in a planned window. Snapshots and staged rollouts keep the change safe.
You can tell the board something better than “we are patching as fast as we can”: you have contained the exposure and the fix follows on a timetable you control. Treat virtual patching as a bridge to the real patch, and give it an owner and an end date.
Adoption: start before the refresh
Many customers meet software-defined networking for the first time through a VCF engagement, and the aim is a roadmap that lets teams adopt it at their own pace. I encourage activating NSX networking as part of the move to VCF 9, even with no workloads on it yet. Teams get hands-on time and see how it operates, and you already hold the licence for the networking layer, so you buy nothing new. Recent releases make a first step lighter: VCF 9.1.1 lets new Kubernetes and VCF Automation deployments start on your existing physical network.
Migration and planning take time, and a rushed networking refresh puts teams under pressure. Early experience also gives the business the evidence it needs for the next step. vDefend then arrives as an add-on at renewal, or ahead of the physical refresh, and existing hardware can run until it ages out. Segmentation can also start on the network you have today.
The foundation for what comes next
The same networking foundation supports outcomes well beyond this blog, and each deserves its own piece:
- Self-service private cloud, where application teams provision their own networks and security within guardrails the infrastructure team sets.
- Compliance and audit support, with segmentation policy you can show an auditor.
- Disaster recovery and stretched networking between sites.
- Cyber recovery, where vDefend isolates an air-gapped clean room and VMware Live Recovery orchestrates the restore.
- Kubernetes through vSphere Kubernetes Service (VKS), with VMware Avi Load Balancer integrating with the networking layer.
- Private AI workloads running on that same networking and security base.
You can do some of this without NSX, at the cost of more manual configuration and more operational overhead.
Businesses care about the outcomes and the services running on the platform. Simpler operations, faster response times and tighter control answer the board’s “so what?”
For anyone weighing a networking refresh, or looking for a first step towards a simpler, faster security position on VCF, Xtravirt’s VCF architects can build a roadmap for your estate. Contact us to talk about VCF and NSX architecture.