Learn Veeam - Core Concepts & Main Architecture

#Introduction

After understanding why Veeam exists in episode 1, now we dissect how it works. Veeam's architecture is like a logistics company: a headquarters that coordinates (Backup Server), a truck fleet that moves goods (Backup Proxies), and warehouses where goods are stored (Repositories). Understanding who does what will save you from confusion when a job fails or performance drops.

In episode 2 we map out Veeam's core components, the transport modes that determine the data path, and supporting components such as the WAN Accelerator, Tape Server, and Enterprise Manager.

#Veeam Architecture: End-to-End Data Flow

#Veeam Backup Server (The Brain)

The Veeam Backup Server is a Windows server that runs all management services: job scheduling, backup cataloging, authentication, and control of other components. It stores metadata (catalog) but does not (by default) move backup data — that is the proxy's job. This is an important distinction: moving data through the management server is a common architectural mistake in other backup products; Veeam separates the two by design.

#Backup Proxies (The Transport Fleet)

The Backup Proxy is the component responsible for moving data between the hypervisor/storage and the repository. VM data is read from the source, processed (dedup & compression — episode 6), then sent to the repository. A single job can use one or many proxies in parallel. Proxies can be Windows or Linux, and can be placed close to the source or close to the repository for efficiency.

Check registered proxies from PowerShell:

Get-VBRBackupProxy | Select-Object Name,HostName,Status | Format-Table -AutoSize

#Repositories (The Warehouse)

The Backup Repository is the storage destination: a Windows/Linux folder, an SMB/NFS share, or S3 object storage. This is where .vbk, .vib, and .vbm files are stored — we dig into storage details in episode 6, including Scale-Out Backup Repository (SOBR) in episode 18.

#A Simple Flow

Hypervisor (VMware/Hyper-V/Proxmox)
        │  VM data read via API / snapshot
        ▼
Backup Proxy  ── process (dedup + compression) ──▶  Backup Repository
        ▲
        │  job commands & metadata
Veeam Backup Server (management + catalog)

#Transport Modes: Data Paths from the Hypervisor

The proxy must read VM data from the hypervisor. How it reads determines performance and network design:

#Direct SAN Access

The proxy reads VMs directly from a SAN LUN (block storage) without going through the production network — the fastest path and the least disruptive to production. The requirement is that the proxy has access to the same LUN (iSCSI/FC zoning). Suitable for ESXi.

#Network Block Device (NBD)

Data is read from the hypervisor over the network using the NBD protocol. Most flexible (no storage access required), but consumes network bandwidth. Suitable for ESXi without SAN access or when the proxy cannot do HotAdd.

#HotAdd

The proxy is a VM running on the same hypervisor. The proxy's disk is hot-added to the source VM while the backup runs — leveraging the virtual storage path without touching the production network. Most common in vSphere environments.

#Hyper-V, Proxmox, and Object Storage

• Hyper-V: Veeam uses Hyper-V WMI sessions or on-host backup.
• Proxmox VE: v13 introduces a data mover on the Proxmox node — a small component that handles native data transfer (episode 4).
• Object storage: direct communication via the S3 API to the repository (episode 6).

#Supporting Architectural Components

#WAN Accelerator

The WAN Accelerator speeds up cross-site backup copies and replication over slow WAN connections using deduplication-aware and compression techniques. It works in pairs: one at the source site, one at the target site. We use it when building off-site copies in episode 7 and DR in episode 9.

#Tape Server

The Tape Server is the bridge to tape libraries (LTO) — still vital in industries that demand long-term archive or compliance. Veeam supports tape-out from repositories and direct backup to tape jobs. We touch on this in episode 19.

#Veeam Backup Enterprise Manager (EM)

Enterprise Manager is a web portal for managing many VBR servers from one place: running restores, viewing reports, and centrally managing licensing. For organizations with multiple backup servers, EM is the unified entry point — we discuss it in episode 20.

#Veeam Backup Server in Lab vs Production

In a lab, a single Windows server can run everything at once (server + proxy + repository). In production, separate the roles: a management server, distributed proxies, dedicated repositories. We break down scaling and hardening in episodes 13 and 18.

#Closing

Key takeaways:

• The VBR Server manages and stores the catalog; the Proxy moves data; the Repository stores the results.
• Transport modes: SAN (direct from storage), NBD (over the network), HotAdd (virtual proxy), plus a dedicated Proxmox data mover.
• The WAN Accelerator speeds up cross-site transfer; Tape Server for archive; EM for multi-server.
• A lab can be all-in-one; production must separate the server/proxy/repository roles.

In the next episode, episode 3, we will install VBR and its components — preparing Windows Server and SQL (built-in Express/single instance), deploying VBR v13, automatically adding proxy and repository, installing per-workload licenses, and preparing credentials and a scale-out repository. The first time you open the Veeam console!