SteelEye DataKeeper Extends Windows Server 2008
The most basic configuration consists of a two node Microsoft failover cluster, with one node in the primary data center and the second node in a different data center at the other side of a WAN link. SteelEye DataKeeper replicates the volume resource(s) defined on the primary server to the secondary server. The volume resource can be a local attached disk, or it can be an iSCSI or fiber channel SAN attached volume. Changes that occur to the volume(s) on the disaster recovery server are tracked in the bitmap and are replicated back to the primary server once it becomes available again.
Stand out features of SteelEye DataKeeper include:
- Host-based synchronous or asynchronous block level volume replication
- Built in WAN optimization enables SteelEye DataKeeper to “pack the pipe” of a high speed/high latency network connection without the need for WAN accelerators
- Compression algorithms make effi cient use of available bandwidth
- Bitmap-based intent logging eliminates the need for complete resyncs
- Bandwidth-throttling provides complete control over replication traffic
- Intuitive wizard-driven MMC 3.0 User Interface
- Supports multiple targets across LAN/WAN for cascading failover
SteelEye DatKeeper Product Brief
DataKeeper Key Features
Block Level Replication SteelEye
DataKeeper replicates data at the block level, ensuring that data is replicated as quickly and efficiently as possible. Implemented as a Windows filter driver, SteelEye DataKeeper sits between the file system and the data volume. By sitting below the file system, open files, locked files, NTFS permissions, encrypted fi les, etc. pose no problem. SteelEye DataKeeper simply sees these files as blocks of data that need to be replicated, just like any other block of data. By removing the need to deal with these issues, SteelEye DataKeeper is able to go about its primary job of real-time replication without adding additional overhead to the host system.
Synchronous or Asynchronous
With two modes of replication, SteelEye DataKeeper is able to offer the most flexibility when choosing the optimal data protection configuration. While choosing synchronous replication guarantees that the target system is always in sync with the source, there is a trade off; the source system must wait for each write to complete on the target volume before it can complete on the source volume. This delay is particularly noticeable on high latency networks and is therefore only recommended on high speed connections with minimal latency, or where absolute data protection exceeds the importance of application performance. Asynchronous replication, on the other hand, allows writes to complete on the source volume while at the same time sending the write to the target system. SteelEye DataKeeper ensures write order integrity by employing a queue on the source system in time of excessive disk writes. The benefit of asynchronous replication is that it allows replication across high-latency WAN links without impacting the performance of the source server. The downside is that in an unexpected failure of the source system, any data in the queue will not be available on the target system.
When pushing data across a high latency network, one must deal with the effect that latency has on the ability to use the available bandwidth. This is especially true once you introduce a combination of high speed and high latency on the same link. SteelEye DataKeeper has been built from the ground up with this requirement in mind – it must not be affected by latency. So, regardless of the latency, SteelEye DataKeeper is able to use up to 90% of the available bandwidth without the use of hardware based WAN accelerators. In addition, SteelEye DataKeeper offers nine levels of compression, which in some cases results in a 5:1 compression ratio, allowing even more efficient use of the available bandwidth.
Challenges and Advantages
|Data Replication Challenges
||SteelEye DataKeeper Advantages
|Ensuring your data is available 24x7x365 regardless of the type of failure: storage subsystem, user error, or site loss
||Provides continuous real-time host-based block-level replication of data insuring multiple copies of your data is available at all times
|Overcoming bandwidth and latency limitations associated with WAN replication
||Optimized for WAN replication
|Avoiding full resynchronizations of critical data
||Incremental synchronization of large database and email repositories
|Replicating transactional email and database data
||Block based replication ensures that regardless of the data type, replication is fast and efficient
|Protecting data using your existing hardware and infrastructure
||Replicate from SAN to SAN, Local Attached to Local Attached, SAN to Local Attached, etc.
|Choosing the correct type of replication for your requirements: asynchronous vs. synchronous
||Choose between synchronous or asynchronous replication, depending upon your data protection requirements
|Completing backups within shrinking backup windows
||Access to the target system allows backups to complete on the replicated data set, allowing backups to complete without impacting production servers