Avamar: how it works

Total data deduplication at source

At source: data are deduplicated before being transferred over the network.

At target: data deduplication is coordinated across the various sites and servers, over time.

Data segmentation: the deduplication process is optimised through the use of small data segments of variable length.

Avamar

Client software agents identify the subfile data segments at source

The backup application only transmits new, unique segments over the network and sends these to a peripheral storage device

Benefits :

• Improved backup windows
• Fewer virtual infrastructure constraints
• Less client-server bandwidth required for the backup process

Simple example of total data deduplication at source

Although results vary according to the combination and type of data concerned, these figures show that Avamar substantially improves backup efficiency and performance.

Avamar's innovative architecture

• RAIN architecture (Redundant Array of Independent Nodes)
      - Each server node has an internal disk drive storage facility and a CPU
      -  The nodes feature high availability are are tolerant to faults

• Grid computing architecture for high performance and online scalability
• Daily Avamar server integrity checks and data restoration capacity checks
• PRAID protection against disk drive faults
• Available for the Avamar software and Avamar Data Store

Avamar's centralised management

Multi-site, rule-based management using existing network bandwidth

• Centralised management of several Avamar systemsar
• Intuitive web interface
• Control panel that provides an instant overview
• Capacity reports and alertss
• Deployment based on one-time universal settings in small companies
• Improves remote/branch office management efficiency and streamlining

Avamar: unrivalled added value

Reduces the volume of data transferred, eliminating bottlenecks, with single-step restoration

• Data deduplication at source :
  – Reduces daily network bandwidth requirements.

• Reduces backup windows :
  – Reduces full daily backup times by reducing the volume of data transferred.

• Centralised management :
  – Streamlines the administration and management of remote backups.

• VMware backup.

• Improves server consolidation by reducing infrastructure-related constraints.

• Single-step restoration :
  – Restores full backups, directories or files in a single step (no more full and incremental backup procedures).

• Guaranteed restoration capacity :
  – Daily integrity checks, RAIN technology and replication, eliminating single points of failure.

Avamar's flexible deployment

Avamar control of office computers/laptops

• Places the power of Avamar in the user's hands :
  – The technical assistance department is no longer required to restore data to office computers/laptops,
  – Data can be restored at any time, from anywhere
  – Automatic restoration management,
  – Document search engine.

• All the benefits of Avamar backup :
  – Data deduplication at source,
  – Full daily backups up to 10 times faster,
  – Minimal impact on the PC and network resources during backup.

Avamar grid replication

Definition

The Avamar solution includes internal data replication technology. This is used to replicate, partially or fully, the configuration and backups from one Avamar grid to another. This replication can be single- or twin-directional. The target may be an active or dormant server. The solution enables users to filter clients and backup strategies prior to replication.

Effective data transfers:

Replication takes place using the IP protocol in asynchronous transfer mode. Upper and lower bandwidth use thresholds can be set based on the company's business hours. Avamar uses data deduplication technology to reduce the amount of bandwidth used and optimise the use of available storage space.

Disaster recovery :

Replication can also be used to copy data stored in one Avamar node (source) to another Avamar node (target). If an incident occurs that affects the first node, the backed up data can then be restored.

How it works

Normal replication

Replication takes place between a source node and a target node. A "replicate" domain is automatically created on the target node during the first replication process. This domain contains an identical mirror of the Avamar source node. The data in the "replicate" domain are read-only. Normal replication occurs automatically at predetermined, programmed intervals.

Only the following operations are authorised:

• Redirect the restoration to other clients which are not included in the "replicate" domain
• Change the backup expiry date
• Backup other clients which are not included in the "replicate" domain
• Delete a backup

Capacities and restrictions

• Only static data are replicated. Each replication transfers all static data from the source node. All operations currently in progress on the source node (e.g. backup in progress, add a new user, update datasets, etc.) will not be replicated. Replication will take place when the data are in "static" mode on the source node.
• During replication, only those data segments that do not exist on the target node will be transferred,
• Backups can be replicated by storage type,
• The Avamar Enterprise Management multi-system management console is recommended for multi-node environments,
• This Avamar technology solution eliminates the use of tapes for disaster recovery plans by providing a simple, cost-effective way of outsourcing daily backup tasks. The scheduler can be used to set replication times and the allotted replication time period.