- Introduction
- Technology Options for Disaster Recovery Solutions
- Quick Checklist for Deployments
- Campus Cluster Maximum Distances
- Campus Cluster Topologies and Components
- Campus Cluster Configurations
- Performance in a Campus Cluster Environment
- Management Aspects of Campus Clusters
- Glossary
- Related Resources
Like this article? We recommend
Like this article? We recommend
Like this article? We recommend
Technology Options for Disaster Recovery Solutions
Campus cluster solutions can form the base for comprehensive disaster recovery solutions that include any one or a combination of the following technologies:
Backup and recovery – Backup scripts and management software automatically generate copies of data regularly, which are then archived safely on site or at a remote location.
Outsourced data services – Some enterprises choose to use an outside vendor to provide a replica of a production data center that can be installed quickly in case of a disaster.
Database replication – Databases may be configured to replicate data to a remote server over an IP network. In case of a failure, a manual procedure would put the remote database into production.
Log shipping – Included in most modern database products, this technology uses the logs produced by the database to "recover" a standby database at a remote site. The remote database is in a standby state, and the logs are applied to it either immediately or after a time gap to prevent logical errors from migrating into the remote database. Logs are usually sent via an IP network, but could also be replicated synchronously using other technologies.
Data replication over networks using Sun StorEdge™ Availability Suite and Sun StorEdge™ Instant Image – Sun StorEdge Availability Suite replicates arbitrary storage volumes to remote sites over IP networks, while the Sun StorEdge Instant Image product allows a point-in-time copy or snapshot to be used. The combination of these two products can be used to replicate a snapshot of a database to a remote site.
Data replication (mirroring) using Fibre Channel (FC) – High-end storage products offer the capability to replicate data to a remote storage subsystem of the same type using direct connections without affecting the servers attached to the storage. This replication usually copies data blocks to a remote site. Typical examples of this type of product are EMC SRDF and Hitachi Data Systems TrueCopy.
Each of these options is adequate for certain situations, but most enterprises choose a combination of several technologies to deploy a complete disaster recovery solution. Each option must be evaluated on how well it meets an enterprise's requirements, its short-term (deployment) and long-term (management) costs, and the level of protection it provides. TABLE 1 compares some of the capabilities of these options.
TABLE 1 Comparison of Disaster Recovery Technologies
|
Technology |
Independent Data |
Clustered |
Automatic Recovery |
Recovery Time |
Maximum Distance2 |
| Remote Backup and Restore | Yes | No | No | High | Network |
| Log Shipping | Yes | No, but possible | No, but possible | Medium | Network |
| Sun StorEdge Availability Suite/Sun StorEdge Instant Image | No | No | No | Medium | Network |
| Storage-Based Replication | No | No | No | Medium | Storage Interconnect or Network |
| Database Replication | Yes | No | No | Medium | Network |
| Campus Clusters | No | Yes | Yes | Fast | 10 km3 |
| Remote Backup and Restore | Yes | No | No | High | Network |
One of the main differences between each of these options is the mechanism used to replicate data. Having two or more copies of data in sync at any time is an advantage for fast, automated failover. However, this approach carries a risk: any defect that exists in the data is mirrored to the other copy, making both copies corrupted or useless. In this case, having a independent copy of data, such as with backup and restore, is essential.
A common alternative is to send logical data packets, such as database logs, to the remote site and apply them to the database after time has elapsed. This approach makes it possible to detect errors—logical, administrative errors as well as hardware related inconsistencies—and prevent them from being applied to the remote copy.
Instead of sending database logs via an IP network, database files can be replicated using Sun StorEdge Availability Suite, either working directly on live data or on a snapshot that could be made using Sun StorEdge Instant Image technology.
Regardless of which combination of options an enterprise may choose, the technologies must be supported by rational processes implemented through careful planning and training.