How to make your Disaster Recovery Architecture distance-proof
For true Disaster Recovery protection, it is essential to mirror critical systems over large distances between primary and backup site. Even with the low probability of a complete outage, the subsequent dramatic impacts of data loss and system (un)availability does require DR prevention measures. Specific requirements include that the application is available with current data even after a complete outage of the primary data center.
In SAP environments for example, most new implementations require Disaster Recovery as part of the architecture just the same as functionality or response times.
Only few customers rely on tape/backup restore as primary method and opt for a ‚hot standby concept‘. Why should there be different standards for new SAP implementations than for existing?
Tape backup and restore is simply not doing the job for true Disaster Protection. When designing a hot standby site, the distance between primary and secondary system is one critical aspect of the architecture. Are two rooms in the same building sufficient? Or is it required to mirror across hundreds of miles?
We always recommend the latter since we have to assume that incidents will affect complete areas. Distance in itself is more a technical and less a conceptual hurdle. Most customers agree on the notion that the further away the secondary system is from the primary, the better. Only traditional block/storage-based mirroring technologies and subsequent high investments lead to smaller and smaller distances.
While vendors try to enhance block-based mirroring with asynchronous mirroring, FibreChannel over IP or others fixes, the concept of block-replication was always designed to mirror within short distances only.
Libelle is mirroring application landscapes in Wide Area Network settings for almost a decade. Key challenges we solved is to address limited bandwidth, network latency, large distance, data integrity and more. The following table gives an overview on the key issue and how Libelle covers them with the patented BusinessShadow solution:
| Problem | Consequeces | Libelle Solution |
|---|---|---|
| Limited Bandwidth | High ongoing fee for network pipe | • Log- and file based |
Unstable Networks | • Block-based mirror • Half-automated log | Automatic resumption |
Network Latency | Latency leads to literal „data jams“ with extended distances | Parallel shipping of data packets using own, |
Libelle "Option Long Distance": Cost-Effective Mirrors in WAN settings
With Libelle „Option Long Distance“, Libelle provides a special edition for Wide Area Networks to provide true distance independence between primary and secondary site.
Read more about our Option Long Distance here...