This Site Is Depreciated
  • This is the depreciated Version of Veeam Backup & Replication Best Practices
  • Contacting Veeam Software
  • DNS Resolution
  • Veeam Backup Server
    • Deployment Method
    • Backup Server Placement
    • Sizing and System Requirements
    • Veeam Backup & Replication Database
    • Protecting Veeam Backup & Replication Configuration
  • Veeam Enterprise Manager
    • vCloud Director Self Service Portal
  • Search Server and Indexing
  • Proxy Servers - Introduction
    • Proxy - VMware vSphere
      • Transport Modes
        • Direct Storage Access
        • Virtual Appliance Mode
        • Network Mode
        • Backup from Storage Snapshots
          • NetApp Data ONTAP integration
          • Nimble Storage integration
        • Selecting a Transport Mode
      • Sizing a Backup Proxy
    • Proxy - Microsoft Hyper-V
    • Proxy - Nutanix AHV
  • Primary and secondary Storage BPs
    • HPE 3PAR VMs disks considerations
    • DellEMC Data Domain advanced scalability
  • Backup Repository
    • Repository Types
      • SMB
      • Deduplication Appliances
      • Integration specifics
      • Windows Server Deduplication
      • Object Storage
    • Repository Planning
      • Sizing
      • Per VM Backup Files
      • Scale-out Backup Repository
        • Capacity Tier
    • Repository HA
  • WAN Acceleration
    • Analyzing Wan Acceleration Workload
    • Comparing WAN Acceleration Modes
    • Sizing For WAN Acceleration
    • Sizing Targets for WAN Accereration Relationship
    • Deployments For WAN Acceleration
    • Is WAN Acceleration Right For me
  • Tape Support
    • Tape Deployments
    • Tape Media Information
    • Tape Config Requirements
    • Tape Parallel Processing
    • Tape Virtual Full
    • Tape Writing to Tape
    • Tape Restores
  • Veeam Explorers
  • Interaction with VMware vSphere
  • Interaction with Microsoft Hyper-V
  • Job Configuration
    • Backup Methods
    • Encryption
    • Deduplication and Compression
    • Backup Job
    • Backup Copy Job
    • Replication Job
    • Application-Aware Image Processing
  • Data Verification Using Virtual Labs
  • Overview of Applications Support
    • Active Directory
    • Microsoft Exchange
    • Microsoft SQL Server
    • Microsoft SharePoint Server
    • Oracle Database
    • MySQL
    • IBM Notes/Domino
    • SAP HANA
  • POC Guide
    • Assessment
    • Accelerated Evaluation
    • Enhanced Evaluation
      • Workshop Example
      • Preparation
      • Automation
  • Infrastructure Hardening
    • Segmentation using Zones
    • Hardening Backup Repository - Linux
    • Hardening Backup Repository - Windows
  • Backup & Replication Anatomy
    • Backup
    • VM Restore
    • Instant VM Recovery
    • Windows File-Level Restore
    • Replication
  • Networking Diagrams
    • Backup Server
    • Proxy Server
    • Repository Server
    • Storage Integration
    • Data Validation
  • Application-aware Image Processing
  • Enterprise Manager
  • Sizing Summary
Powered by GitBook
On this page

Was this helpful?

  1. WAN Acceleration

Deployments For WAN Acceleration

PreviousSizing Targets for WAN Accereration RelationshipNextIs WAN Acceleration Right For me

Last updated 4 years ago

Was this helpful?

As the source WAN accelerator can only process one task at a time (one VM disk in a backup copy job or replication job), you may need to deploy multiple WAN accelerator pairs to meet the performance demands.

As the target WAN accelerator can handle multiple incoming streams (as described in the section of the User Guide), it is recommended to maintain a 4:1 ratio between the number of source WAN accelerators per target WAN accelerator.

This guideline is very much dependent on the WAN link speed. Many source sites with low bandwidth will create little pressure on the target WAN accelerator. So, for instance, in multiple ROBO configurations a 10:1 ratio can be considered.

If there are sites with very high bandwidth (such as datacenter-to-datacenter replication), they will produce a much more significant load on both the target WAN accelerator and the target repository due to the second data block lookup (for more information, refer to the ).

Note: The secondary data block lookup is used, when a data block is not available in the WAN accelerator cache. When there is a WAN cache “miss”, the secondary lookup for the same data block is performed on the target repository. If it is found here, it is read back to the WAN accelerator instead of re-transmitting over WAN.

Assuming the source and target repositories can deliver the throughput required for the optimal processing rate, use the guidelines that follow.

Note: The numbers below are processing rates. The WAN link usage is dependent on the achieved data reduction ratio.

  • Average throughput per target WAN accelerator: 500 Mbit/s (62.5 MB/s)

  • Depending on the achieved data reduction rate (typically 10x), the transfer rate over the WAN link will vary.

    • If the processing rate is 62.5 MB/s, and the data reduction rate is 10x, then it is possible to sustain 6.25 MB/s (50 Mbit/s) over the WAN link.

    • If the WAN link has high bandwidth (above 100Mbps) consider using backup copy jobs without WAN Acceleration. However, if you use WAN accelerators in that scenario, it may require deployment of multiple WAN accelerator to fully saturate the WAN link.

Many-to-One WAN Acceleration
User Guide