Transcript
On prem applications and data require an on prem backup solution. Myth or reality? What do you think? Well, so where did this one come from? Where do we think this myth came from? And you need to sort of look back at the history of the way we did data protection 20 years ago. And a lot of these sort of myths, I think, come from, um, people's experiences from, you know, ten, 20 years ago. And we had a situation where we had relatively slow networks to what we have today. You know, we may not have had an external connection into the, um, into the internet or externally from our data center like we did in the we do today. And as a result, it was a sort of a requirement to have your backup data as close as possible to the system you're going to restore, because networking was expensive and relatively slow. Now, exclude the fact that if you've got a D.R. site, you'd still have to be putting that data protection system in another location. We'll come back to that in a second. So I think that myth has come from from that scenario. Um, and, and obviously the one thing that people think about all the time is that when we do a restore, they think it's an all or nothing thing, that when we restore, we have to restore our everything. And actually, if you look at the different styles of failure that occur in a data center, you can fat finger data and delete it by accident. Somebody could delete it. Um, because there's a bug and a bit of software, you might have a minor hardware failure, or you might have an entire data center go down. And it could be any or any or all of those sorts of things that that cause you to have a problem. And as a result, you might only be restoring a small amount of data for a certain requirement. So why do you need big pipes to restore, you know, masses of data? The assumption is that you're restoring everything, and that's not necessarily true. Now, even if you are restoring everything, you need to think, well, what scenario am I doing that in my data center might have gone well. There's no point having an on prem copy in the same data center, but you've now lost, so you've still got to put a copy somewhere. You can't keep it on on prem. There's still an issue there. So I think, you know, the idea of needing to have an on premises copy in order to restore is not true, because actually it's more complicated than than it seems. And I, you know, I don't know what your view on that one is, Matt, but, you know, the architecture of the way you designed the backup, you simply just wouldn't want that to be the case anyway. Oh I mean a couple of really great great points there. You know, I think it it does all come from that old ideology that you had smallpipes. And so you had to, to really have kind of that locality in terms of, uh, protection and recovery. You know, I can't move that data very fast, very far. So I'd better have it close, you know, uh, which led to, you know, a lot of the synchronous, asynchronous kind of replication mods at different layers so that you had, you know, I, I've got another copy over there, so I don't have to move it or recover it over there. It's already there, you know, which of course goes to our previous slide about just adding cost, expense and complexity in there. Um, but the more important point, I think, is the nature of the disasters we're trying to build around and architect for recovery for. To your point, you know, it's very rare, uh, that you're seeing like an entire site going down. You know, you may see an application go down, you may see, you know, a subsystem or, you know, important network get affected or, you know, more likely, you know, in the event of kind of a cyber incident, you'll see parts of things all over the place, uh, that are affected. So you can't necessarily just smash everything back in place because then you'd be, uh, overwriting the good data along with the bad.
