 Hi, TCO Sopil Bhartiya, and welcome to T3, a more topic of this month. The topic of this month is data. And today we have with us once again, Dave Burmingham, director of customer success at Sios Technology. Dave, it's great to have you on the show. Thanks for having me. Yeah, I'm glad to be here talking with you today. Yeah, it's my pleasure to host you again. And today, when we look at, you know, of course data is a big, you know, we can talk about some aspect. But what we are going to talk about today specifically is high availability in the cloud. How would you define high availability or disaster recovery in the cloud native world? Because sometimes when we look at cloud, it does take care of a lot of things. Availability is something they offer. Of course, recovery from any failure is also they offer. But sometimes people folks that cloud is a magical place and you don't really need. So I want to understand high availability, disaster recovery in the cloud native world. First off, you know, high availability and disaster recovery are key aspects of system management that, you know, aim to ensure that your services are not disrupted and that data is not lost in the event of a problem. In the cloud native world, HA is typically achieved through redundancy and failover mechanisms with services running on multiple instances across different zones and different regions in the cloud. Disaster recovery, on the other hand, can involve things such as backing up systems and data and generally having a plan in place to restore systems from those backups in the event of a disaster such as a system-wide outage. Now running in the cloud, you mentioned, you know, it's a magical place and that is kind of, you know, some people might think that, right? So, oh, I'm moving to the cloud. I no longer have to worry about high availability and disaster recovery. And of course, you know, cloud vendors do provide several tools and mechanisms to help and facilitate high availability and disaster recovery, but it's really ultimately up to the customer to properly utilize these in their architecture. And, you know, the cloud is not immune to outages and errors and other issues. We've seen, you know, major outages, minor outages, you know, it's just as susceptible to outages as your own data center. So, you know, therefore you must still have a robust high availability and disaster recovery plan in place to assure uninterrupted service and data protection. As we talked about cloud, you know, a lot of times, depending on who you talk to when we look at cloud, a lot of time folks thinks that since everything is running in the cloud, you really don't have to worry about high availability or disaster recovery. Folks also feel that practices like chaos and ring and resiliency and, you know, redundancy that kind of eliminates the need of having, you know, any HADR either approach or solution. But the fact is that these two words have to coexist. One is more about practices and culture and other is more about actual solutions and services, whatever you want to call them. So I do want to talk a bit about these two things as well. Yeah, and that is a common misconception I hear all the time, you know, people moving to the cloud and, you know, we don't need to have, you know, we don't need to do anything in terms of backups or, you know, doing anything to ensure resiliency. And, you know, they're really two different things. You know, at Psyos, we help in both areas, but let's take a look at the, you know, the resiliency or I think of the, you know, the high availability aspect. So, you know, being able to keep your data in sync in a synchronous fashion. So no data loss across multiple cloud zones, availability zones or, you know, every cloud vendor has a different name for it. But getting it from one data center to another and having a duplicate copy in real time is absolutely what you need to do for high availability. That's for the data protection. And then of course the application monitoring and failover that ensures that you survive from, you know, minor blips or user error or something like that. So you can have, you know, the four nines of availability, very minimal downtime, you know, five minutes or less a month of downtime. But that's, if you eliminate that, that's great. And you should absolutely do that. But that does not eliminate the need for having a robust backup strategy and a disaster recovery strategy. You may lose the entire region. So it doesn't matter that you have your data copied synchronously between availability zones in that region. That entire region could go offline, you know, could be natural disaster, could be human error, whatever this, and we've seen that happen. So you need to have that disaster recovery plan for that major type outages, but maybe even more likely you, you know, you're hearing the news of people, you know, how these people coming in and they're holding your data hostage. You're encrypting all your data and they want you to send them, you know, a million dollars in Bitcoin and you know, they're gonna give you the key to unencrypt your data. Either you pay the, you know, you send them the Bitcoin or you hope that you have an archive copy that is offsite inaccessible, something that the bad actors have not gotten access to to encrypt that as well. So you still need to have that failsafe or that airbag in place for those situations you just can't expect to happen. So that's no different than on-prem in the cloud. It's the same. The cloud just gives you some additional tools, you know, a lot of the cloud vendors have backup and so that, you know, you can take images of your systems and replicate them, you know, and incrementally, you know, maybe with a five minute RTO or 15 minute RPO or thereabouts, and that's great. You know, you should absolutely take advantage of those functionalities. At SIOs, we can do real-time asynchronous replication between regions that gives you even a lower RPO, but that's, you know, you need to take advantage of those tools. And of course, the infrastructure in place in the cloud, where else do you have the ability to have, you know, 20 or more data centers spread across the entire world. So you, you know, how you design your disaster recovery plan has just become a whole lot easier in terms of the infrastructure you don't need to build out, you know, data centers just to host your disaster recovery. Cloud vendor's already done that for you. When we look at not only just SIOs as ecosystem, but in general, there are companies who are in different stages of their cloud native or digital transformation journey. A lot of companies, of course, they were born in the cloud era, but a lot of companies which are, had been around for a while, which means they do have a lot of experience of HNDR in on-prem, but as they migrate, they move to the cloud. Of course, they have to look at it from the lens of cloud. So talk a bit about how is HNDR different in cloud as compared to on-prem, and how do you folks enable these users to kind of seamlessly continue their journey? High availability and disaster recovery is different in the cloud compared to on-prem. For a few reasons. So first, the cloud platforms, like I was mentioning earlier, they provide native tools and services for HNDR such as automated backups and replication of the storage on the backend, plus the easy setup of redundant instances with that cross-region replication. So that can make HNDR easier to implement in the cloud compared to on-prem because all those tools and utilities and infrastructure are already in place. Well, however, with the cloud, there's a greater emphasis on automation and infrastructure as code, which can change how your high availability and disaster recovery strategies are designed and implemented. And then you have to think about the shared responsibility model of the cloud, which means that while the cloud provider takes care of certain aspects of availability, like hardware redundancy and network redundancy, storage redundancy, the customer is responsible for everything else, like the application layer and making sure that your data is highly available. And that's what we do. We have the tools and the clustering software LifeKeeper that's going to monitor the applications here and take recovery actions. And then DataKeeper continuously replicating with synchronous block level replication or between zones or asynchronous between regions and then even enabling hybrid cloud models. So people either migrating to the cloud, looking for a way to get their data into the cloud or just running in a hybrid model on a regular basis with maybe the cloud being their disaster recovery center instead of managing their own. You know, they still have their own data center, but they're using the cloud as their DR site. So at PSYOS, we can help facilitate that with our HA and disaster recovery solutions. When you're talking about how different HADR is in cloud as compared to on-prem, do you also see your customers' users see some unique challenges that they face when it's come to HADR in the cloud? Once again, VisaVis on-prem. You know, often people are new to the cloud. So there are some nuances. Depending upon the cloud, there's different, you know, way networks, virtual networks are configured that it's gonna be a learning curve. So one of the challenges is just really understanding and properly using all the high availability and disaster recovery tools and just the infrastructure tools in general provided by the cloud platforms. So each cloud vendor has its own unique set of tools and services and I can be a learning curve. And if you're doing multi-cloud, then that's two times because you know, every cloud vendor is gonna have different tools and different ways to do things. Another issue, you know, the unique challenges with the cloud is just knowing, you know, where is my data? You know, data sovereignty and compliance, that's something that everyone needs to be concerned about. And some, you know, some regions have very specific laws regarding where your data can be stored and transferred. And so that will certainly affect how you implement high availability and disaster recovery in the cloud. And then, you know, finally costs can be a concern. So, you know, cloud providers again offer some utilities for high availability and disaster recovery but everything comes with a price tag and it's usually, you know, consumption based model. So often, you know, using them is going to incur additional charges getting a good estimate and managing that on an ongoing basis can be a challenge with, you know, with creep, you know, things, you know, being implemented in the cloud. And all of a sudden, your bill, kind of like my cable bills seem to go up a couple of dollars every month. And so managing that is certainly a challenge. Since you mentioned cost, I also want to talk a bit about the egress fee that is associated when you migrate or move data and organization, they do have to move data for so many different reasons. How does kind of Sios help organizations with that? So moving data at Sios is with our product data keeper and data keeper runs on Windows and Linux and it does real time block level replication from, you know, one server to another from one to multiple from on-prem into the cloud or between instances and different availability zones in the same region or even between regions with our asynchronous replication. And so data keeper will be obviously moving data when we're talking about from region to region or zone to zone, there's going to be egress data cost associated with that. So Sios helps manage that by and allowing you to enable data compression before the data leaves the region or the, you know, the availability zone will compress it and then decompress it once it's received. So that can save, you know, say when your egress data costs depending upon the data type of me, I've seen, you know, three to one, four to one, five to one compression ratios. And so that obviously will have a direct impact on your data egress costs. What advice do you have for customers so that they can build great HHA strategy for cloud? I'd say first you have to understand your needs. So not every application or service may require the same level of availability. So all this comes down into your, you know, your business continuity plan. You need to categorize your applications. Which one needs real-time replication? Which one is, you know, maybe backup, maybe, you know, 15-minute RTO, RTO, you have to define what that is. So prioritize according to your business needs. It's not a one-size-fits-all model. So that's number one. Number two is design for failure. You have to assume that every component can fail. And so you must design your architecture in a way that such failures have minimal impact on the overall service availability. I mean, we've seen, I remember a large outage, you know, Microsoft had with, I think it was Azure Active Directory, one offline in one region. And that had such a wide impact across many of their services that they were offering. So you have to look very closely at, you know, at the end of the line, you have this service that someone's consuming, but there are so many components along that chain that need to be available in order for that end service to function. So you need to definitely design for failure for each and every one of those components. The third recommendation I would say is automate where possible. We talked about cloud automation and infrastructure as a code, the automation can significantly reduce the risk of human error and increase the efficiency and speed up your recovery time. So, you know, one of the things you could do for the, you know, the high availability is making sure that you have the clustering tools in place, like, you know, like the science lifekeeper that will continuously monitor your applications and automatically take the recovery action when needed. Or even if it's not automatic if you're doing disaster recovery and you wanna push the button, being able to recover all your applications with one single push is so much better than having to, you know, have your engineer who you don't know at any time of a disaster who's gonna be available to implement all the steps in a long run book. If you have the step is push the button that's so much better than, you know, having to go through a 20 page run book to recover an application. And then finally, you have to test. So, you can have all the best plans in place in the world and you think you've covered everything but if you haven't tested it and regularly test it then you're doing yourself a disservice. So, regular testing ensures that your plan works as expected and it will help you discover any potential issues. So, make sure you schedule your disaster recovery and your HA test on a regular basis. Dave, thank you so much for taking time out today and of course talk about HA in the cloud. Thanks for all those insights and as usual, I would love to chat with you again soon. Thank you. Yeah, I'm looking forward to it. See you next time. Thank you.