EMC Avamar - Architecture

1. Overview

EMC Avamar is a comprehensive, client-server network backup and restore solution. With its unique global data deduplication technology, Avamar addresses the data protection challenges in today’s IT environments.

• The ever-increasing amount of data to backup presents a challenge to organizations facing the demands of shorter backup windows, quicker restore responses, consistent backups of remote sites, and regulatory requirements; all with the need to accomplish this with fewer staff and tighter budgets.

• Avamar meets these challenges by re-designing backup and restore as true disk-based processes. Avamar’s patented global deduplication technology reduces the amount of backup data by identifying unique data at the source. Avamar stores only one copy of this common data across the backup network. This results in a dramatic reduction in the amount of data that is moved across the network and stored in backup storage. The same data is backed up as in traditional backup systems, but consumes significantly less network and backup resources as only unique data is stored. And, by using standard IP network technologies, dedicated backup networks are not required.

• Avamar employs a scalable disk-based, server architecture built of modules that provide a balance of connectivity, security, processing and disk storage resources. Scheduled backup and replication functionality enable efficient backup of remote sites and provide disaster recovery of primary backup sites. Avamar provides a user-friendly interface for central management of the entire backup system.

Tradional Backup

A high percentage of data that is retained on backup media by most backup solutions is highly redundant. The typical backup process for most organizations consists of a series of daily incremental backups and weekly full backups.

• Daily backups are usually retained for a few weeks and weekly full backups are retained for several months to several years. Because of this process, multiple copies of identical or slowly-changing data are retained on backup media, leading to a high level of data redundancy.

• A large number of operating systems, application files and data files are common across multiple systems in an enterprise. Identical files such as Word documents, PowerPoint presentations and Excel spreadsheets, are stored by many users across an environment. Backups of these systems contain a large number of identical files.

• Additionally, many users keep multiple versions of files that they are currently working on. Many of these files differ only slightly from other versions, but are seen by backup applications as new data that must be protected.

• Backing up redundant data increases the amount of backup storage needed and can negatively impact network bandwidth. Organizations are running out of backup window time and facing difficulties meeting recovery objectives due to the need to manage backup versions and a myriad of backup tapes.

Avamar Advantage

Avamar differs from traditional backup and restore solutions by identifying and storing only unique, sub-file data objects. Redundant data is identified at the source, drastically reducing the amount of backup data that travels across the network to be stored and managed by the backup host. When storing data objects, Avamar takes maximum advantage of inherent hard-disk characteristics. Avamar also creates and stores “trees” that link all data objects from a single backup. These “trees” are used to re-create files for restore.

2. Features

• Global data deduplication ensures that data objects are only backed up once across the backup environment.

• Systematic fault tolerance, using RAID, RAIN, checkpoints and replication, provides data integrity and disaster recovery protection

• Highly reliable, inexpensive disk storage is used for primary backup storage.

• Since standard IP network technologies optimize the use of the network for backup, dedicated backup networks are not required. Daily full backups are possible using existing networks and infrastructure.

• Scalable server architecture provides security and expandability. Additional storage nodes can be added to an Avamar multi-node server to accommodate increased backup storage requirements.

• Flexible deployment options include Avamar Virtual Edition and Avamar Data Store. Avamar supports a wide-variety of client operating systems and applications, including: Windows, Linux, Unix, NDMP, Microsoft SQL, Microsoft Exchange, SharePoint, and Oracle. With its global deduplication technology, Avamar is an efficient backup choice for VMware and remote office backup environments.

• Centralized management is also provided. Avamar Enterprise Manager and Avamar Administrator interfaces enable remote management and monitoring of Avamar servers from a centralized location via internet access. Avamar can also integrate with Data Protection Advisor and Backup & Recovery Manager for further monitoring capabilities.

Terminology

• An object is a single instance of deduplicated data. Objects are stored and managed within stripes on the Avamar server. An object is also sometimes referred to as a chunk.

• A stripe is a unit of disk drive space managed by Avamar. Objects are stored within data stripes.

• A node is a self-contained, rack-mountable network-addressable computer consisting of both processing power and hard drive storage. Nodes run Avamar server software on the Linux operating system.

• A server is a group of one or more nodes on a local, high-speed network.

• A system is one or more Avamar servers and the servers or desktop clients that back up data to those Avamar servers.

Avamar COmponents

• The Avamar Server stores client backups and provides essential processes and services required for client access and remote system administration. Avamar Administrator Server (mcs) and Avamar Data Server (gsan) run on the Avamar server.

• Avamar Client software runs on each computer or network server that is being backed up. Avamar provides client software for various computing platforms. Each client consists of a client agent and one or more plug-ins.

• Avamar Administrator is a user management console software application that is used to remotely administer an Avamar system from a supported Windows or Linux computer

Avamar Server Node Types

• Utility nodes are dedicated to providing internal Avamar server processes and services, including the administrator server, external authentication, Network Time Protocol (NTP), and web access. 
• Storage nodes include the Avamar Data Server software and are dedicated to providing backup storage. 
• The NDMP Accelerator is an optional specialized node that, when used as part of an Avamar system, provides a complete backup and recovery solution for NAS devices via the Network Data Management Protocol (NDMP). Avamar supports EMC Isilon, VNX, and Celerra and NetApp filers with the NDMP Accelerator. 
• The Media Access Node is an optional node that can be used as a pass-through device for sending Avamar backup data to tape for long term storage

Systematic Fault Tolerance

• RAID (redundant array of independent disks) is a method of protection for disk data corruption. RAID is a balance between performance and efficiency. Avamar servers are protected by either RAID-1 or RAID-6, depending on the configuration. Avamar also has hot-swap capability with minimum system impact for highest failure-rate components (more than 90% of expected failures). 
• RAIN (redundant array of independent nodes) provides failover and fault tolerance across nodes. RAIN provides uninterrupted functionality during node failure, replacement and reconstruction. In the unlikely event of a node failure, the backup data will be stored on the remaining nodes; data for recoveries is reconstructed using parity. RAIN is used to replace the failed node, reconstruct the data on the replacement node, and when expanding an Avamar server, rebalance the capacity across all nodes. 
• Replication protects against data loss in the event of a server loss. Efficient, scheduled replication (local or remote) ensures availability/redundancy of data if primary server is lost. 
• Checkpoints protect the server in the event of operational failures. They provide redundancy across time. Checkpoints are a read-only snapshot of the Avamar server taken to facilitate server rollbacks. They are created using hard-links to all the stripes. Regular checkpoint validation, including auto-repair capability, is used to ensure data integrity. 
• High Availability Uplink and Dual Switches provide high availability in the event of hardware failure.