Replication helps protect your data and files by producing a duplicate copy at a second site, server, or storage array. I covered host-based replication in a previous blog.

In this blog, I’ll cover two other types of replication — array-based replication and network (or fabric) based replication.

Array-based replication

Array-based replication requires a central data storage unit (SAN or NAS) and a partner unit. With array-based replication, the SAN or NAS processes the data and the commands to process and validate the data being replicated.

Advantages of array-based replication

  • The work is offloaded from the servers to the storage device.
  • You only need one location to control many replications of multiple servers.
  • Hosts (Servers) are not required at the second site or to be attached to the second SAN/NAS.
  • A central SQL server can be set up to replicate with the servers that actually present applications to users, such as order tracking applications.
  • The right software can queue databases to ensure that transactions and the database are in a recoverable state.

Disadvantages of array-based replication

  • Cost per device can be high, especially when you’re not replicating all of the data on the SAN.
  • Only SAN or NAS based data can be replicated or controlled.
  • A second SAN or NAS is required, increasing the cost for the solution.
  • There could be compatibility problems of replication technology/software between SAN/NAS hardware and vendors.

Examples of array-based replication software

Network-based replication

The last type of replication is network (or fabric) based replication. This type of replication works separately from the hosts (servers) and the storage devices. A device on the network intercepts packets being sent to and from hosts and arrays and copies them. These copies are replicated to a second device that then replays the packets at a second location. The devices are, in essence, splitters. The data goes in and then it’s split out to different sources.

Advantages of network-based replication

  • It’s a separate component from the SAN/NAS or the hosts.
  • Processing is independent to the host and SAN/NAS.
  • It allows replication between multi-vendor products.

Disadvantages of network-based replication

  • The cost of implementing devices to support this kind of replication is high.
  • Newer technology for the data center, standards, and process are still being worked out.
  • There are a limited number of “players” in this area of replication.

Examples of network-based replication products