One of the most common reasons I see for unacceptable recovery times is a reliance on cloud based backup. Pushing data off site every night is easy, especially in modern paradigms where we only have to send out the changed bits rather than a full copy of the local data. The problem with cloud backup becomes apparent when you try to bring data back in house for a rehearsal or a recovery.
We call this “recovery lag”. There are two challenges you face when recovering from a third party backup provider. The first is a potential the second is a definite.
Challenge 1: How quickly does your backup system create a restore set?
You might expect a backup solution to have a restore point immediately available, This isn’t actually the case; In some cloud backup solutions an initial seed of data is taken and from that day onwards only changed bits are transmitted during the backup process. This technology is great for reducing the amount if data you need to send off site every night, (we normally base our calculations on 5% to 15% of the total estate). However when you need a full Restore Set you have to take that initial seed data and then apply every single nightly backup to create a current restore point. Technologies like Veeam handle this well by regularly scheduling a full backup to reset the seed. Many other technologies don’t.
So what does this mean? In some technologies if your original seed is 200 days old, and you want to restore, then your backup system has to take the seed and apply 199 updates to it in order before it becomes a workable restore set. This can take time.
Challenge 2: Is your restore set easily transferable to the restore location?
Let’s presume you are an SME, you have a 50mb connection to the internet and your backup partner, A raid array failed over the weekend, It’s 8:00am and your backup restore set is now ready for you pull back and it’s 4TB in size. It’s going to take 25 hours of watching progress bars before you can even start recovering servers.
You can probably see why we call this recovery Lag.
We do things a bit differently because we focus on recovery rather than just getting data off site.:
Our Hybrid Cloud offers full failover into the cloud and also local failover onto hardware onsite with your production equipment; The big difference here is that we store each recovery point as a full data set. Meaning that you can immediately mount that restore point and use it. Either as a Bare Metal Restore source or as a mounted virtual machine.
Our Cloud based backup is hosted at our own data centres on the same network infrastructure as our server recovery suites and our hardware warehouse. This gives a client options:
1. We can recover your data to one of our servers and ship it to you “ready to boot”
2. You could run your full recovery project in one of our server recovery suites
Next Steps? You can talk to me or any of the team here at Adam Continuity / Centerprise. We can have a full and frank discussion around what your stakeholders expect, where you budget lies and then we can introduce one of our dedicated solutions team to help build out a Continuity Solution that fits your requirement.