REQUEST A QUOTE
Data Replication Technologies
As we’ve said elsewhere, replication is backup. You never know when a natural or manmade disaster will mean you need to restore. As we’ve also said elsewhere, it isn’t quite that simple.
Data replication technology can be synchronous or asynchronous. The difference is simple: in synchronous replication, data is written to both sites (primary and back-up, or secondary) at the same time; in asynchronous replication, there’s a delay in writing to the backup.
One might imagine that that is a no-brainer, and that all replication should be synchronous so that data on both sites is identical at all times – but, once again, it isn’t as simple as that. Synchronous application costs more than asynchronous; more importantly, it brings with it latency that slows the primary application, and it can’t be used over distances less than about 30 miles or greater than about 180 miles.
So, which to choose? (Bearing in mind that this decision will have a bearing on how far away the back-up site can be, and, therefore, where it can be). Where the recovery time objective (RTO) is low – you must be able to get it back at a speed that is close to instantaneous – and/or the possibility of data loss is unthinkable, synchronous replication is probably the one you want. For data recovery, asynchronous replication may be better because:
- It works over long distances; and
- It needs less bandwidth.
Never forget, though, that in asynchronous replication the back-up data will not be quite up to date with the source data, and so there is a risk that a system outage may cause data loss.
If this sounds like a subject for discussion with an expert, that’s probably because it is. We’re ready to talk it through any time you say the word.