Master Data Backup: Host, LAN, Server-less & NDMP Techniques Explained

Preface

The article introduces basic backup concepts using a carousel format, aiming to consolidate knowledge of backup software by analyzing the advantages and disadvantages of various backup technologies.

Evolution of Backup

Backup copies all or part of data from application hosts to other storage media to prevent data loss caused by errors or failures. Main networking models include Host, LAN, LAN-free, Server-free, and Server-less.

1. Host Backup

The original method copies data directly from the production server to backup media. It offers fast transfer and simple management but lacks sharing capabilities, making it unsuitable for large-scale backups.

Comment: Host backup is like copying C: to D:, fast but not shareable for modern large data needs.

2. LAN Backup

All production systems install backup clients and connect to a backup server via network. The server controls backup and transfers data over the network to the media.

Comment: LAN backup solves sharing issues of Host backup but heavily depends on network and server resources, leading to possible congestion.

3. LAN-free Backup

Production systems have clients and media management software; the backup server only sends commands, and the systems transfer data directly to the media, reducing server load.

Comment: LAN-free frees network resources and reduces server pressure, but production systems still bear load; Server-free later addresses this.

4. Server-free Backup

Combines array snapshots: a snapshot is created at the production data point, presented to the backup server, which mounts and copies the snapshot to media. Snapshot creation is fast, resulting in near‑zero backup window and no impact on production systems.

Comment: Server-free eliminates server resource usage, prompting the need for a completely server‑less approach.

5. Server-less Backup

Data moves directly between source storage and backup media (e.g., storage system to tape library) without server involvement, using SCSI‑3 Extend Copy or NDMP protocols, offering high performance and low resource consumption.

Comment: SCSI‑3 requires high‑end storage and lacks flexibility; NDMP, a TCP/IP‑based standard for NAS backup, is widely adopted.

Main Backup Software Overview

EMC and Symantec provide multiple backup products for different scenarios.

EMC offers NetWorker (software) and Avamar (appliance or virtual edition).

Symantec offers NetBackup and Backup Exec , both software.

Commvault provides Simpana ; IBM offers Tivoli Storage Manager (now IBM Spectrum Protect).

EMC Backup Software

1. NetWorker

Acquired from Legato in 2003, NetWorker replaced EMC Data Manager and became EMC’s core backup solution. It supports backup, archiving, and NDMP for NAS. Data flows from clients to storage servers/nodes, then to media such as disks, tape libraries, or network shares.

2. Avamar

Acquired in 2006, Avamar focuses on data deduplication, often used with NetWorker. It comes as an appliance or virtual edition (AVE) that can use third‑party storage (SAN, NAS, DAS). In virtual mode, storage is allocated from the host VMware ESXi server.

Symantec Backup Software

1. NetBackup (NBU)

A cross‑platform, high‑end product supporting Windows, Linux, Unix, and many databases. It offers backup and archiving, supports LAN‑FREE and SAN‑FREE, and uses a three‑tier architecture: Clients/Agents, Master Server, Media Servers.

2. Backup Exec (BE)

Targeted at Windows, low‑end market. Supports Windows and Linux clients, limited database support (mostly Oracle on Windows). Offers file backup and LAN‑FREE.

Commvault Backup Software

Simpana

Provides data protection, archiving, replication, resource management, and search modules. Architecture includes CommServe (control), Media Agent, and iDataAgent. Data moves from iDataAgent to Media Agent, then to media (disk, tape, network share).

IBM Backup Software

Known as TSM (Tivoli Storage Manager), now IBM Spectrum Protect. Offers enterprise‑grade backup, archiving, space management, and disaster recovery. Uses an Agent‑Server model similar to BE.

Backup Software Feature Analysis

Backup vs. Archiving

Backup stores data for recovery and retains multiple versions; archiving moves infrequently used data to cheaper storage for retrieval.

Symantec BE provides only backup; archiving requires Enterprise Vault. NetWorker, Simpana, NBU, and TSM include both backup and archiving.

Data Deduplication

Deduplication reduces redundant data by replacing duplicate blocks with pointers.

EMC NetWorker lacks built‑in deduplication; it requires Avamar or Data Domain. Simpana, BE, NBU, and TSM have built‑in deduplication.

NDMP Backup

NetWorker, Simpana, BE, NBU, and TSM all support NDMP, offering generic NAS NDMP configuration without vendor lock‑in.

OS Compatibility

TSM, Simpana, and NBU have the most comprehensive OS support; NetWorker follows; BE has the least.

Maintainability

TSM has a complex installation and is hard to operate; other products are easier to use. Overall, functional differences are minor, with Symantec offering flexible pricing and product tiers, while others are more uniform.

NDMP Technology and NAS Backup

NDMP was co‑developed by NetApp and Legato (acquired by EMC) in 1996, with version V4 widely adopted; it is supported by most major NAS vendors.

Traditional NAS Backup

Because NAS devices are closed systems, backup clients run on production servers, and data is read via CIFS/NFS and sent over the network to backup servers, consuming network and server resources.

Comment: Traditional method burdens network and server resources, limiting performance for large data volumes.

What NDMP Does

NDMP enables NAS devices to send data directly to attached tape libraries or backup servers without a backup client. It supports three networking modes: 2‑way, 3‑way, and Remote.

NDMP Networking Modes

2‑way

The backup medium connects directly to the NAS; the backup server issues commands, and the NAS writes data to the tape library, improving performance and reducing server load, but the tape library is dedicated to that NAS.

Comment: Direct data flow boosts transfer speed, but tape library is tied to a single NAS.

3‑way

NAS systems send data over a dedicated backup network to a NAS that is attached to the tape library, allowing multiple NAS devices to share the tape library, though performance is limited by network bandwidth.

Comment: 3‑way enables sharing but adds network latency.

Remote

The backup software handles tape operations; the NAS only transfers data to the backup server, reducing NAS requirements and allowing shared tape resources, though performance may suffer and server load increases.

Comment: Remote mode lowers NAS demands and enables tape sharing, but adds network transfer and server burden.

Choosing the appropriate NDMP mode depends on specific scenarios and requirements.