00:00
Hello everyone. I'm vin senior Product manager at Pure storage And today we're going to talk about active workload. If you are part of the storage infrastructure team, you must have had the need to move specific workloads from the storage area to another storage area.
00:25
It could be because with your growing needs you acquire more storage hardware and you want to move certain workloads from the existing storage areas to the new ones or because there are some more demanding workloads now that your existing area cannot accommodate and you need to find the storage area as a destination that has the resources and capacity to accommodate those or your hardware is just at the end of life in it to retire it and to evacuate the
00:55
workloads running on it and move it to a new area. Whatever the reason, maybe you're not alone, there are some existing approaches which you could take to move workloads from an arid weather. However, the problem is that they all lead to disruption to the host and apps running against that storage area.
01:18
For that problem to make workload mobility easier. We have released active workload. Active workload is a non disruptive workload mobility feature that is based off our stretched clustering technology. Our stretch clustering technology is also a foundational block for our high
01:42
availability active cluster solution and now we have enhanced our stretch clustering technology to support not just one but from one array to up to five different arrays connectivity which facilitates this workload mobility use case and makes it much more simpler. Now let me tell you how it works. So in this diagram, as you see, we have a Flasher X, let's assume that you are pure customer and you start with
02:11
the Flasher X which is running three different types of workloads. VM workload, a DB workload and a V D. I workload. Your host is running over here, which has the apps that are being used by the business and your host is connected to this array at this point of time, all the IOS are happening against this Flasher however, you have purchased a brand new
02:37
Flasher Excel for the most demanding workload in your environment, which is the database workload and you decide to move it. What would you do? The step one is to establish a connection between the two arrays your source and your target. So let's draw it out with a dotted line.
02:57
Once you establish the connection, the next step is to stretch this value, which is where our stretching, stretch clustering technology comes in while stretching, establishes a synchronous replication relationship between the tour is for this specific volume and it will start replicating this database workload the volume associated with it to the
03:27
flattery excel. The target. Very once the data has been replicated completely, this volume will show up here now, until this point of time, the host is still sending the IOS to this array and there's no disruption, there is no change in how this was working so far.
03:51
So the businesses don't even know what's happening now, once the data is available on the second flash array, this volume is essentially the same volume in terms of its serial number. So from this flash array you would configure this host to be connected. Let's write a solid line and expose more paths from this host
04:19
to the same volume. Once you can figure that the IOS start happening against this flattery, you can then remove or let the older parts be based on your host and at that point of time you can move this volume and deposit it locally to this flash and that's it.
04:48
You connect the target flash array, you stretch the volume using the part 3rd, you deposit the volume into that flash array throughout this process. Initially the host was using this flash array and now the host is using this flash array. We are different path however, there's no destruction to the host and the apps. Now, this is very useful if you just want to move workloads from a sorcery to a target flash
05:17
array. What happens if you have a fleet of flash arrays and you want to move this workload to this array and then this workload to some other array or the video workload to the next day? For that reason we have, like I said, enhanced our uh stretch clustering technology to support multiple flash array connection. So while this is happening you could also
05:42
establish a connection with the dotted line, let's say to the second flash array in your fleet and then stretch this volume as well and start replicating it once it shows up on the second flash array, you will do the same host connection from here to here and then allow the app and the host to forward the IOS onto
06:18
the second flash. With that you can move not just two workloads simultaneously, but we actually can support up to 12, a toad, 1/4 And the 5th flattery with this Array itself and you can move workloads across
06:48
five different flash arrays simultaneously in one go. Now, whenever I share this kind of functionality with our customers who have been using traditional storage platforms, they always ask me when I like this feature, but tell me how much does it cost. And I always tell them that with pure, you're subscribing to innovation. If europe,
07:10
your customer, you just need to upgrade your software to 640 and above and you can enjoy the benefits of active workload. There's no extra licensing, there's no extra hardware required to enable this feature. So I hope this utility is useful for you. Thanks for tuning in today and for more details, you can click the link below
