Case Study
9 September, 2026
How Bezeq and TeraSky used VCF Automation, VCF Operations Orchestrator, and Infinidat to make a complex restore process dependable, repeatable, and easier to operate.
Provisioning is the visible face of a private cloud. Recovery is the moment that proves whether the platform can be trusted.
Bezeq is one of Israel’s leading telecommunications companies, providing communications infrastructure and services to businesses and consumers nationwide. Its large-scale technology environment supports critical services that require high levels of availability, operational consistency, and control.
At Bezeq, restoring a virtual machine from an enterprise storage snapshot involved far more than selecting a recovery point. The operation crossed storage, virtualization, cloud delivery, validation, and cleanup. It also had to work across an Infinidat environment with multiple endpoint hosts, where the correct storage context had to be identified for each workload.
The underlying technologies already provide everything required to recover a workload. The opportunity was to turn that capability into a dependable platform service – one that preserves expert decisions, reduces coordination effort, and gives operations teams a clear path from request to recovery.
The project was not about automating one storage task. It was about treating the entire recovery path as a product.
Recovery Is a Coordination Challenge
A restore procedure can look straightforward on paper, yet every step carries context. The correct snapshot must match the selected virtual machine. The right Infinidat endpoint and host relationship must be used. The restored datastore must be handled safely. The virtual machine must be prepared before it is introduced into the cloud environment. Temporary resources must then be removed in the correct order.
When those decisions live mainly in runbooks and individual experience, recovery depends on who is available and how familiar that person is with every participating system. Even a technically successful restore can leave behind operational issues if validation, isolation, recording, or cleanup is missed.
We designed the Bezeq workflow around a different idea: the operator should express the intended recovery result, while the platform should own the sequence, checks, and system-to-system coordination required to deliver it.
One Guided Request, One Accountable Flow
VCF Automation provides the service entry point, while VCF Operations Orchestrator carries the request through the infrastructure landscape. The workflow determines the workload context, locates the appropriate recovery source, coordinates the required Infinidat and VMware operations, prepares the restored virtual machine, delivers it into the target cloud environment, and records the result.
The multi-endpoint storage design is handled inside the automation. Rather than asking the operator to identify which Infinidat endpoint or host object a workload belongs to, the workflow resolves that relationship from the workload and datastore context. This removes a fragile decision from the request process and keeps the experience consistent as the environment grows.
The original recovery point remains protected. The workflow uses a temporary writable copy for the restore operation, applies isolation controls to the restored workload, and includes cleanup as part of the same execution. Recovery, therefore, ends with both a usable result and a cleaner infrastructure state.
Business and Operational Advantages
The value of the solution is visible in day-to-day operations, not only in the workflow diagram.
A single recovery path. Storage, virtualization, and cloud-delivery activities are coordinated into a single execution rather than managed as separate handoffs.
Knowledge is retained in the platform. The decisions normally made by experienced engineers are encoded into validations, routing logic, safety checks, and cleanup behavior.
Repeatability across storage endpoints. The same service can resolve the correct Infinidat context for different workloads without exposing endpoint complexity to the operator.
Safer restored workloads. Isolation and identity-related preparation are applied before the recovered machine is made available, reducing the chance of conflicts with the original workload.
Faster diagnosis. The workflow has recognizable stages, execution history, and recorded recovery details, allowing support teams to identify where an issue occurred and what was already completed.
Cleanup with ownership. Temporary snapshots, mappings, and datastores are treated as part of the service lifecycle rather than as optional follow-up work.
More value from existing investments. Bezeq continues to use its established Infinidat and VMware platforms; VCF adds a consistent service layer across them rather than introducing another isolated recovery tool.
Why VCF Mattered During Development
The platform also changed how we built and evolved the solution. VCF Operations Orchestrator provided us with a reusable integration layer for VMware and external systems, enabling the development team to focus on Bezeq’s recovery logic rather than re-implementing basic platform connectivity at every step.
The workflow could be assembled from focused actions, tested in stages, and then exposed via VCF Automation as a single service. That modular approach was important: storage discovery, snapshot handling, vSphere recovery, cloud delivery, validation, and cleanup can evolve independently while the request experience remains stable.
Development and operations also stay connected. The same workflow that performs the recovery provides execution history and clear failure boundaries. This shortens the distance between building an automation and supporting it in production, because operational visibility is part of the platform rather than an afterthought.
A Pattern That Extends Beyond Restore
The most important result is not a single automated procedure. It is a repeatable method for turning specialist infrastructure knowledge into an operational capability that more teams can use safely.
The same pattern can support additional Day-2 services: identify the request context, choose the correct endpoint, apply environment-specific rules, coordinate several systems, verify the outcome, record what happened, and close the lifecycle cleanly. New policies or infrastructure endpoints can be incorporated without asking users to learn a new process each time.
That is a meaningful advantage of building on VMware Cloud Foundation. The platform used to deliver workloads can also include procedures to protect, recover, and manage them throughout their life. Automation becomes part of platform engineering rather than a collection of disconnected scripts.
Final Thoughts
A recovery service earns trust when it is understandable, repeatable, and designed to finish safely. Speed is valuable, but confidence comes from knowing that the right source was selected, the right systems were coordinated, the restored workload was prepared correctly, and temporary resources were not forgotten.
For Bezeq, VCF Automation and VCF Operations Orchestrator provided the foundation to make those expectations executable. Infinidat supplied the enterprise recovery capability; VCF turned the surrounding process into a platform service that is easier to request, operate, observe, and extend.
From my perspective as a Broadcom VCF Knight at TeraSky, this is where platform automation has its greatest impact: not in hiding infrastructure, but in preserving the expertise required to operate it well – and making that expertise available every time the organization needs it.
About the Author
Kobi Vaknin is a Cloud Automation Engineer at TeraSky and a Broadcom VCF Knight. He designs VMware Cloud Foundation automation for complex enterprise environments, focusing on turning cross-platform operational procedures into dependable, reusable services. His work spans VCF Automation, VCF Operations Orchestrator, vSphere, cloud delivery platforms, and third-party infrastructure integrations.