Business Continuity Practices for Diverse Datasets

Sandro Lima: My name Sandro Lima, I’m director of Technical Business Development at DataCore Software. So, today, [we’re] going to talk about how DataCore can help you accommodate a diverse data set on your business continuity practices. So, we have some good content here, so let’s jump right into it. All right. So, there’s no point in emphasizing how important critical data is, right? But what happens is that you cannot treat all data equal, and with equal importance in your business continuity and disaster recovery plans. So, what you need to do is understand what is their potential impact and the risks associated with each data set, so to provide the appropriate levels of protection to each one of them.

And from a data storage perspective, the only way you can do that is having a wide range of capabilities that can give you these appropriate data protections, right? Depending on the RTO and RPO requirements of your data. I’m going to give [unintelligible 0:01:27] details on each on of this – or some of these – capabilities [further], but now I just want to highlight some of the challenges that you may face when you try to implement this wide range of capabilities across your entire storage infrastructure.

Problem: Multiple Storage Vendors

So, the first one, and very common, is having multiple storage vendors. So it’s not uncommon for a company to have two or more different storage [array] vendors, and what happens [is] that each one of them has different capabilities and most of the time they don’t talk to each other, so what happens that even when they have – they can provide the same capability, for example in this example, imagine that you have vendor one, two, and three, they all offer Snapshot, for example, and even then they have different configurations, different interfaces, and you end up having to create three different processes, and treat them separately in your BCDR plan. So, just adds complexity.

Problem: Different Storage Models

So, second challenge [you – very common as well], is different storage models. So, you have companies with a traditional – [leveraging] traditional storage arrays, connected through SANs – Storage Area Network; you have some companies that are more on a [converged] model, leveraging direct attach and then internal storage; some companies doing hyperconverged or very small footprint, simple deployment, but what we see very frequently across our customer base is a hybrid scenario, where companies have all sorts of different models and including cloud for some archival or disaster recovery. So, combining different storage vendors with different storage models, then you can see how this becomes really complex for you to basically deploy all these capabilities across all your storage infrastructure.

But it doesn’t end up there. So, even if you lay out a well-designed, well-tested BCDR plan, what happens that your environment is always changing. So, for all the lifetime of your data center, you have technology refreshed, right? So, you bring in new vendors, new architectures, new technologies. You have to deal with mergers and acquisitions, and then you – suddenly you have to accommodate and ingest other vendors, architectures, [personnel] procedures and everything. You can have data center changes, so either like a physical move or an expansion. Maybe your company wants to leverage cloud more aggressively for replication, for disaster recovery. So – and this list just goes on and on, right? So, [the] point here is things are changing all the time, and you have to – your BCDR plan needs to flexible enough to accommodate these changes over time.

Solution: DataCore

All right. So, with this scenario, so [you] have different vendors, different – different vendors, different storage models, and an ever-changing environment. So, how can you still accomplish our goal of having this wide-range of capabilities consistently available across your entire storage infrastructure? So, at DataCore, so we believe that it all starts with having a very comprehensive set of [advanced] services delivered through a single platform. And this is exactly what DataCore delivers, right? So, we can’t – we have synchronous mirroring, synchronous replication, continuous data protection that [we’re] going to talk in a minute, snapshots and a lot more of this enterprise [created] data services that can be deployed across your storage infrastructure in a very consistent way. And with this platform, you can integrate and ingest any sort of storage direct-connect, SAN, so you can see at the bottom, so we have ISCSI, fiber channel, NVME, direct-attached, cloud, right?

And on the top, where [you] see the consumers, so we can serve this storage to different host types of people – fiscal services, to hypervisors, to virtual machines, and containers. So, that’s the first step – having this wide range of data services available through a single platform. A second step, and as important as the first one, is [a] deployment model, and this is where being [soft-defined] really enables DataCore to be deployed on any model. So, we can deploy DataCore as storage controllers, [for] primary and secondary storage, on a hyper-converged scenario, you can install it on the cloud, on a disaster recovery site – so, doesn’t’ matter what kind of model we have today, or you plan to have in the future, so we will be able to accommodate any vendor that you have now or in the future, you’ll be able to accommodate that with DataCore.

And now we can see a picture of what I was [drawing], so basically having this consistent set of [wide] services, across your entire data center. So, you probably can see now how it can simplify your processes and the design of your business continuity and disaster recovery plans.

DataCore Capabilities

Soo, next, let’s talk about a bit more in detail about some of these capabilities, right? So, the first one is when you’re talking about data, this is very critical, right? So, we’re talking about RTO in the range of microseconds to milliseconds, and zero RPO. So, it means that you can’t have like any data loss, and you have uninterrupted data access. So, for these situations, synchronous mirroring and [unintelligible 0:08:36], is [key], right? So, just talking a little bit more about it, so synchronous mirroring allow you to have two copies of the data on two nodes acting as one. So, as soon as one becomes unavailable, the second one is [ready] to become primary, and avoiding any data interruption. And system makes sure that these datas are kept in sync all the time.

So, this diagram illustrates how this works, so very simple – we have two nodes, three [hosts] on the top, imagine that you have two discs being served, so VD one for example, on the left-hand side, is served primarily by node one, and having node two as secondary. So, they’re both active and in sync, so whenever node one becomes unavailable, so node two can pick up and in the host, we see no interruption on his data access. So, another example of synchronous mirroring is what we call stretched or [meta]-clusters. So, it’s basically, you have these two nodes not on the same place but several miles apart in different locations still. So, as long as you have a low latency link connecting both, you can have all benefits of synchronous mirroring, so no data loss; no data interruption, but with an added redundancy, because being on separate locations mean that you have independent [unintelligible 0:10:25] resources, independent calling, so benefits of synchronous mirroring with added redundancy. So, that would be a second scenario. So, synchronous mirroring with auto-[fillover] would be ideal for your most frequent data that needs to be available all the time and accept no data loss.

Third example would be having a third copy of your data. So, in this case, you have your two nodes on an HA [fashion], and then you can have a third copy of the data, but this would be on a standby [fashion], right? So, this becomes very handy, whenever you have to take one note out for maintenance, for example, but your data is so critical that you can’t lose the HA access to your data. So, what you can do in this case is you bring down node one for maintenance for example, and node three can be manually brought up, and then it will be an HA pair for node two. When node one is back it just like fell back to node one, and everything goes back to normal. So, this gives you the flexibility of doing these maintenance windows without affecting your HA for your data.

All right, moving on. So, second model of – second model is when your RTOs are in the range of seconds to minutes, but [you see] you have strict RPOs, so in this case, [async] replication will be a good fit. So, just to explain a little bit more of [what] synchronous replication is the ability for you to have the primary disk is streaming all the changes across network for a remote location, right? So, this picture here illustrates the most common scenario for synchronous replication, this is to build DR sites. So, you see on the left you have two nodes in an HA fashion, so synchronous mirroring between them, so HA, and all the changes are asynchronous, replicated to the DR site. And whenever something happens with the primary data center, so the second copy [at] the DR site, you can bring it up and then you can recover from that. Another example, very common as well, of using replication, is to use for secondary storage. So, sometimes you want to have a separate copy of your data, but you don’t want to store this on your fast and expensive primary storage, so you can use asynchronous replication to basically stream all these changes to a secondary, slower but more cheaper storage.

All right, so moving on. Our third case is when you have RTOs between minutes and hours to even days, right? And in this case, snapshots and backups are commonplace. So – but a particular case is when you still want this not-so-strict RTOs but you have very strict RPOs, and in this case, CDP – that stands for continuous data protection – can be a really good fit. So, let me explain a little bit more about CDP. So, CDP allows you to have constant recording and log of all the changes that happens in your disks. So, a common analogy that we make is with a DVR – so digital video recorder – so imagine that you are recording everything that is going on live, and you are able to rewind to any point in time and create an image at this point.

So, a very interesting example – and it’s becoming more and more popular unfortunately, is a protection against ransomware. So, imagine that – so CDP allows you to protect these critical disks and if there’s a ransomware outbreak, you can basically rewind to the exact second before the event, and be able to recover from that. So, [you] basically have near to zero data loss and can recover very quickly. Another thing that we see customers doing is not only creating an image for the recovery, but also creating other images over time to – basically for forensics and analysis, basically investigation and trying to understand how the infection took place in the first place. But that’s not the only use case for CDP. We also have, for example, protection against accidental deletion of files as well.

All right. And the last one is snapshots. So, snapshots very common so everybody knows snapshots, so it allows you to capture this point in time, image of your disk. So – but there are some added benefits when you basic – you take snapshots with DataCore. So, the first benefit is that you’re doing it at the storage level. So, you can take snapshots on the – at the host level, but this depends on [whole software] and it takes resources out of the host, right? So, doing that at the storage, we can do it completely transparent for the host. So, that’s the first one. So, the second one is that DataCore gives you the ability to create [differential] snapshots. These are very lightweight snapshots that you can take. So, if you want to take snapshots very frequently, so these differential snapshots are a good option.

And third, if you want a full snapshot, or a full copy or image of the disk, you can also do that, and DataCore offers you the ability to store this snapshot on any other storage. So, you don’t need to, for example if you’re taking a snapshot of [your] very hot data, you don’t need to store that and consume resources on this fast and expensive storage. So, you can move it to another storage that can be on another vendor, so [then] store that – the snapshots there.

So, this is what I have for today, so if you’re interested in learning more, we have a business continuity page on our website, so it’s lots of resources and [I] also want to emphasize that, in addition to these [few] capabilities that we covered today, there’s a lot more, especially on the performance and efficiency arena, so head to [our] DataCore website and look for this link and you’re going to see like explanations and details on all features of DataCore.