Overview of Backup Architecture, Features, and Data Protection Strategies

Starting today, I will gradually share a series of knowledge about backup architecture and features. Data protection solutions are mainly divided into two categories: backup and disaster recovery. Backup provides multiple redundant copies of data so that logical errors such as accidental deletion, virus infection, or loss can be recovered from the backup copies, ensuring little or no data loss. Continuous Data Protection (CDP) is also a real‑time backup technology.

Disaster recovery, on the other hand, is a technology that ensures business continuity. The primary site and the disaster‑recovery site keep data essentially identical, so errors at the primary site are also replicated to the disaster site. When a logical or physical failure occurs, disaster‑recovery switching quickly brings the service back online, guaranteeing continuous operation.

Backup Technology Overview

There are many backup methods, such as local snapshots, data replication, cloning, imaging, database RMAN backups, and virtual‑machine backup functions. However, these techniques usually protect only server or storage‑local data, and backup and recovery strategies rely heavily on manual administrator actions, lacking effective lifecycle management. Backup software can solve these problems and supports various network backup modes (LanBase, LanFree, ServerFree, etc.) to meet changing business needs.

Today the backup software market is crowded and competitive. Products targeting virtualized environments include Veeam, NetVault, eBackup; general‑purpose backup solutions (physical and virtual) include Veritas NBU, Simpana, Data Protector, Legato Networker, TSM, Time Navigator, Attix5, Anybackup, DBackup, etc. Integrated hardware‑software backup appliances include Avamar, Veritas appliances, Huawei HDP, AceSure/BoxSure, and others.

Backup Software Architecture

Traditional backup software typically adopts a two‑tier architecture: a backup management server and a media server. The management server handles backup tasks, policy creation, scheduling, metadata, and client information, providing GUI/WEB interfaces. The media server writes client data to the actual backup media. Different vendors use different names (e.g., Simpana calls the management server CommVault Server, Veritas calls it Master Server; media servers are called MediaAgent, MediaServer, or Proxy Server), but their roles are the same. Simpana is a leader in Gartner’s Magic Quadrant, so we use it as an example.

In Simpana, the backup management server is called CS (CommVault Server), the media server is MA (Media Agent), and the backup client installed on business servers is iDA (iDataAgent). Under CS control, iDA reads data from the client and transfers it to MA, which then writes it to the chosen backup medium. iDA comes in many types, including file, database, and mail system agents.

Compared with other backup products, Simpana is an enterprise‑grade data management platform that integrates backup, archiving, and data management functions, solving management difficulty and data sharing problems. Its Content Store is the core of policy management and virtual storage resource pooling, providing automatic tiered storage, deduplication, encryption, and intelligent indexing. The CS, MA, and backup media libraries are all components of the Content Store.

CommCell is the backup management domain of CommVault. Within a CommCell, all backup resources, plans, policies, data, and users can be shared. A CommCell contains a single backup management server that governs the domain, but it can include multiple media servers and multiple iDA agents to achieve concurrent backup. Details about backup networking and reliability will be covered in later articles.

Multi‑Backup Domain Management

Simpana supports centralized reporting and management of multiple CommVault domains, consolidating several branch domains into a primary backup domain where reports can be viewed centrally.

Each branch domain runs independently (each has its own CS and MA). Network failures between the primary and branch domains do not affect each other. Branch domains automatically push metadata to the primary domain, and when the network connection is restored, data transmission is automatically repaired.

Simpana also provides multi‑domain tiered data protection. Branch domains deduplicate and replicate data (Copy1) to the primary domain’s central repository (Copy2). The primary domain has multiple MA instances: one for central backup and another for interaction with branch‑site MAs. Data is copied between MAs, enabling centralized backup and management across domains.

Tiered protection can copy the centrally backed‑up data from the primary domain to disk, tape, or cloud storage, creating a third copy (Copy3) for long‑term offline retention. During recovery, the console in the primary domain can restore data regardless of whether the copy resides locally or in remote offline storage.

Data Archiving and Recovery

Simpana also integrates data archiving, which retains only the data copy while deleting the original to save storage space. It supports three archiving methods: OnePass backup‑archiving integration, regulatory‑compliant mail archiving, and traditional lifecycle‑based tiered storage archiving.

Typically backup and archiving are separate modules (e.g., Veritas NBU for backup, Enterprise Vault for archiving). Simpana unifies them on a single platform. OnePass combines backup, archiving, and reporting by sharing indexes and performing a single data collection and movement, eliminating the need to scan, index, and move the same data multiple times.

Simpana mail compliance is designed according to regulatory standards and supports Exchange and IBM Domino. It monitors all inbound and outbound mail in real time, copies each mail to a log mailbox, assigns a sequence number and timestamp, creates content and audit indexes, and finally writes the archived mail to a WORM device.

The compliance feature provides authorized users with search, review, retention, and result‑export capabilities, preventing modification or loss of archived data and reducing security risks.

Traditional archiving moves data based on its “temperature” to different backup media, usually combined with backup: data is first backed up according to policy, then less‑frequently accessed data is moved to near‑line or offline storage (often tape libraries) for long‑term preservation and easy retrieval.

Both backup and archiving must ensure data accessibility. Stub files provide transparent callbacks to data stored on different media, allowing users and applications to access data without being affected by the underlying storage location.

Differences Between Backup and Archiving

Backup primarily creates data copies for recovery after damage or loss, meeting RTO and RPO requirements; the copies are full or incremental replicas of the original data. Archiving, however, is intended for long‑term preservation to satisfy regulatory compliance or audit needs; typically only one copy is kept, the original data is deleted, and a stub remains in primary storage to retrieve the archived data when needed.

The “ICT Architect Alliance” gathers experts from various technical fields and provides a point‑to‑point, real‑name communication channel for precise exchange and assistance. To protect personal information, only both parties who exchange business cards can see each other's details. Interested experts are welcome to click “Read Original” to join.

Friendly Reminder: Please search for “ICT_Architect” or scan the QR code below to follow the public account and get more exciting content.