Video: Modernizing Database Resilience: The Shift to Automated, RSC UI-Based Recovery | Duration: 2708s | Summary: Modernizing Database Resilience: The Shift to Automated, RSC UI-Based Recovery | Chapters: Introduction to Recovery (26.619999s), Data Recovery Essentials (198.81s), Legacy Tools Challenges (324.865s), Flexible Recovery Automation (616.11s), Database Recovery Demo (821.54s), Sensitive Data Discovery (1392.35s), Sensitive Data Discovery (1568.745s), Conclusion and Validation (1994.595s)
Transcript for "Modernizing Database Resilience: The Shift to Automated, RSC UI-Based Recovery":
Hello, and welcome to Database Recovery Without the Panic. A walk through of Rubrik's database protection and how we enable and simplify recovery. Before we get started, I always wanna take a few moments to just kinda set the ground rules. We're gonna talk a lot about features and functions around the Rubrik platform today. There's always some consideration as we share this information with you. Just like I said, setting ground rules as we get into this. As we get going, just a brief introduction. Myself, my name's Mike Fall. I am database practice lead here at Rubrik, basically helping customers understand and focus on how Rubrik protects databases. And supporting me today is Alpika Singh. She's a senior technical marketing manager and basically does a lot of work around building materials for sharing out Rubrik information. The agenda for today is pretty straightforward. We're gonna be covering recovery and really how Rubric can simplify that for you, streamline it, make your life easier. And with that, we always wanna understand the beginning, why is it that having that kind of tool set, that kind of functionality is important? Because we're always when a recovery happens and we're in the, you know, the firing line of a disaster, we gotta make sure we know how to recover, how to recover quickly. And so there's always a bit of panic that we always go through. Certainly, I did when I was a DBA when I was starting my when I had to do a fast recovery. And with those legacy tools and the complexity, which we'll get into, there's always those challenges. Now as we go through that, right, addressing the panic moment, addressing how Rubrik can address that, We'll dig a little deeper into our recovery options. And again, there'll be a demo on this as we get into things. And then once we get through all that, we wanna talk a little bit about our sensitive data discovery feature, as well as a live demo for that. So getting started, right? The whole recovery panic, right? Like, why is it such a big deal? Certainly, when we start talking about recoveries and the amount of data that we're protecting, these are just a few numbers from our research, right, where we can see data is exploding. We're never managing less data. We're always managing more. And really, the reality is we're managing more with kind of the same number of people, right? There's a lot of enterprises. They're exploding on their data estate, but they've still got the same number of people to manage it to address that. So really, there's this kind of vice grip on data professionals as we grow the environment, but not add to the headcount to do that, to manage that with. Of course, there's a reality cost to what happens when the data is not available, right? If there's a disaster, it's estimated that, you know, the average cost per minute, per minute, not hour, not day, but per minute, is $9,000 Obviously, that can vary depending on the enterprise, but that's a lot of money that even if you're down for a few minutes can really impact the business. And then kind of folding into that, right, we see more and more customers are going to experience some sort of attack from a cyber event, right? Again, 90% of customers in the last year, at least based on our research, shows that they've been impacted in some way or another. And so that's downtime, that's money, right? That's stuff that we have to be able to recover quickly from. Of course, that starts to ask the question of, well, what's important to us then when we have to figure out our recovery, right? So obviously, we need to have our recovery predictable. Know, one of the things when I was a DBA, I'd always be testing my restore processes, right? And the more that I test that, the more I'm comfortable with it, the more predictable it is, the more I know that when I'm under the gun, I can do it and I know how long it's gonna take me. Of course, what's equally important is the integrity of that data. It doesn't do me any good to bring up a corrupted copy of the database or have, you know, recovered data that has been affected or compromised. So really making sure that we have a clean copy of the data as part of our recovery is vital. You know, it has to be fast. That's sort of the third thing there. Right? I've never had anybody complain to me that a restore was too slow. I'm sorry. It was a restore was too fast. You know, it's it's really all about how quickly can we bring the data back up and again, being clean and predictable. And then obviously what folds into all this is making it simple, right? The more complicated tools are, the more we're gonna impact our predictability and our ability to do this quickly, our speed. So the simpler we can get it, the more streamlined we can get it, the better it's gonna be overall. Now we think about the legacy tools, right? And that's where we start to run into challenges, particularly with those items that we just addressed on the previous slide, right? A lot of times we lose predictability because there's a bit of a guessing game as to where data lies. Right? We write backups off to a storage location that gets copied off to somewhere else. So now we might have multiple teams trying to figure out where all the files are. Certainly, if I have to assemble a long string of files to do a recovery with, then I might be having to figure out, okay, do I have all the files? Am I missing any? Are there any gaps to my recovery chain? Right? And as we think about this, right, and having to guess at all the where the data lives, how it's being managed, that adds to a lot of manual work that has to be done in order to do a recovery. Right? And that will slow us down. That will hamper us. That kind of comes to the last piece or the next piece is that slow data movement, right? It's okay. We have to assemble everything. We may have to copy things from one location to another. Then we have to restore it as part of the database restore. And those are sort of laws of physics problems. But again, the more work I have to do copying bits and bytes across the wire, the more that that's gonna impact my recovery time. And then finally, there was always a day where I shouldn't say always, but there was a time where it's like, oh, well, we only had managed maybe one, maybe two databases. Right? And I didn't really have to worry about like different backup methods. You know, nowadays, enterprises are gonna have multiple database platforms. You might have SQL. You might have Postgres. You might have Mongo. Right? You might have four or five different database platforms all getting back up in a different way. And so now you have to context switch across all those database platforms in order to execute a recovery. So it's a lot of work, a lot of things that the legacy solutions sort of trap you into. And again, these are things that Rubrik is looking to address. And really, when we start talking about the Rubrik advantages, it's really going after all those those kind of trap situations. Right? First off, we wanna build a lot of automation. It says automated discovery, but really, it's more than that. It's automated discovery. It's automated protection. It's doing more with less. How can Rubrik simplify the whole process by automating as much of it as possible for you? The next step there or the next piece is this policy driven approach. Right? Traditionally, with the legacy tools, we would set up backup schedules. Right? We'd run some backups on Sunday nights and others on Monday nights or Tuesday nights and and basically stagger these all out. And we'd have to be the ones responsible for sorting out when backups could run and when they wouldn't be contending with one another. Rubric basically tries to simplify this by going with this more policy driven approach, where you go in, you declare, this is how many backups I want and how long I wanna keep them for. And then Rubric will go ahead and figure out the backup schedules for you. Right? So we're gonna take that heavy lifting. We're gonna try and run as many backups in parallel as possible, but recognize there's a resource limit. So then we will stagger things out so that we don't have to use up all your resources, choke out the environment. Instant access is a neat feature, what we call a live mount, and we'll see more of this in the demo. But the live mount is the ability to rapidly bring up a copy of your database, right? Again, going back to the exploded data volumes. Right? We have databases that are easily one, five, 10 terabytes in size. And that takes time to recover even if I'm just I've got a streamlined process. Well, what if you had a tool where I could bring up those databases literally in a matter of minutes and access that data quickly? And that's what the Rubrik Live Mount is. Again, we're gonna see more of it in the demo, but really it's it's one of these key features that a lot of our customers really like because they can quickly get a data, say somebody dropped 500 rows from a terabyte database instead of having to fully recover that database. Now I can just live mount it and quickly get at those 500 rows. The last item there, the unified visibility management. Again, the idea is that Rubrik's gonna streamline everything for you, right? We're gonna give you a consistent way to manage your databases through the Rubrik platform, and then you let Rubrik worry about the heavy lifting. Again, all databases are gonna be backed up differently, restored differently. Right? That's just the nature of the beast. But with Rubrik, let Rubrik worry about all those details. You work through the Rubrik interface and you can apply stuff in a consistent fashion. You can run recoveries in a consistent manner. You know, the interface, the workflow's all gonna be the same for you. And that really can simplify your life by reducing that context switching. To dig a little bit more into that backup policy piece, I do wanna touch on what our SLA domain automation is. So really, it's a fairly straightforward policy that says, hey, I'm gonna set a frequency. Maybe I want a one backup a day or one backup every twelve hours or one backup every eight hours. And then how long I wanna retain those? So maybe I'll have daily backups for a month and then roll that into a monthly backup that I keep for a year and then a yearly backup that I keep for three years, right? So you have a lot of flexibility here to sort out how you want the backups to be taken. But what's nice is then Rubrik will take this and actually schedule the backups for you. You don't have to do any of the heavy lifting of making sure backups aren't conflicting with one another, any resource contention. It's all done behind the scenes for you. Now, then we start talking about, okay, how do we leverage this for recovery? And again, the legacy approach would be, I've got this recovery chain that I have to use. Right? I have to assemble all my files, get my my level zeros and my level ones and get all my archive logs or my transaction logs, get them in order. Right? And this was a very manual process. So we look at this, right? And this is a fairly typical way of doing a recovery if I wanna get back to a point in time. Well, Rubrik says, what if I can take all this and simplify it? And that's where we get to our flexible recovery here. Right? The flexible recovery is fairly simple. Right? We again, we're gonna do all the heavy lifting for you. We know how to assemble the files. We know how to put them all in order. We know where all the files live. So all you have to do is come into a timeline view in our UI, select a point in time, and then allow Rubrik to figure out how you get back to that point in time. And this is the same if I'm doing a traditional recovery or a live mount. Again, we're gonna see that in the demo. The last item I wanna touch on is, you know, most DBAs, they're like, uIs are alright, but I'm a command line person. I like to automate. I like to script. Well, Rubrik gives you that functionality as well. Right? While we do a lot of things to the UI, the UI, everything you do in it, is essentially a call to our GraphQL API. And that GraphQL API is exposed to our customers either using software development kits or other scripting methods. And so what you can then do is extend Rubrik. Right? You can use the UI, but you can also write your own scripts, your own automation that wrap around those interfaces, those APIs, and allow you to execute those tasks as part of a broader automation scheme you have. And really what this all does is this takes those those long recovery times. Right? If we think about the legacy and all the different steps that we would have to go through, all the different things we would have to process, You know, recoveries could easily take four to eight hours. Right? Granted, this is a little bit of a a number that's, you know, throwing a dart dartboard just grabbing something because, you know, the reality is is it could be 12. It could be 24. Depending on the scale of your environment and the work that you have to do to assemble your files, it can really extend out those recovery times. But then we go to Rubrik where everything's automated, everything's streamlined. You don't have to context switch between the different platforms. You get the live bound. You can start to access your data literally in a matter of minutes in the event of a disaster. Well, that's enough talking for me. I think that goes through a lot of the features and functionalities, but let's go ahead and look at how this actually works. I'm gonna hand it off to Alpika to do a demo for you. Thank you so much, Mike. Good morning. Good afternoon, everyone, depending on where you're joining us from. So like Mike mentioned, I'm Alpika Singh, and I'm a senior technical marketing manager here at Trupic. What I'm here to do is show you how the recovery actually works and all the options that are available with Rubrik. Mike did a wonderful job in explaining what the options are. Let's take them. Let's go ahead and see them, live. One thing to note, I have a bit of a cold going on, so I might be taking some breaks and pauses just to clear my throat. Here is Rubrik Security Cloud. This is our SaaS based, centralized platform, pretty much that manages all your workloads across cloud, SaaS, and your on prem. So all your databases, different clouds, different SaaS applications. So let's head straight to our Microsoft SQL database to see what we have protected. Now, as you see here, there are a lot of SQL Server databases that's actually being protected. So let's dive into one of the instances. Now, again, like lots of databases, I'm going to pick one of them. That's F1. And what you see here is a calendar view of all the dates for which for which backups have actually taken place and you can perform recovery from, right? So, let's select today, which is the seventeenth. And then what you see here is this timeline view of when the log backups were actually taken. Like Mike mentioned before, you can specify all of this, how frequently you want to back up your data, how frequently you want to back up your logs. All of that granularity is available to you and what you see here is the result of that. So now, if you want to do a recovery, you can actually do the recovery from the snapshot itself or what you can do is just slide the timeline view to the closest available log backup and then just simply select recover. Now, what rubric will do is it will take the snapshot to come from it and apply all the log files automatically for you. So, everything is up to date for you and you don't have to do anything manually. Now, let's take a look at the recovery options. So you do have the option of mount. And Mike mentioned a lot about mount, where we are, where databases can be recovered almost instantly and it can be available for your use. Now, the good point is that the actual underlying storage infrastructure is is actually hosted on two big cluster. That's the reason why the mount is so quick and it's available for you so quickly. The other options include restore where you are replacing the existing database with a copy of the database. So it's like in paste recovery and you do have the option to export the database as well. Now, a lot of times you have test and dev needs where you don't want to obviously test against your live database. So you can create a copy of it and simply select on export and you can select a list of compatible host where you can do the perform data copy. If I just select the host to show you, you know, just a simple advanced export, you can even set up parts, data file export parts. You can set all that up before recovery itself. Now, we will focus a bit on mount, but I just want to highlight the other option. So you do have the option to download all the files and then perform all the recovery yourself. For the mount option, it's fairly simple. You just select the mount option. Click next. Select where you want to mount the actual database. It can be on the same host or it can be on another compatible host as well. Like if your host is down and you want to get your database up and running quickly, you can definitely use one of the existing host and then proceed further. Select the instance you want to mount and then just give it a name. And once you click mount, Rubrik will automatically mount the database on the host. And you can proceed either doing recovery, recovering few files or actually using the database for your other applications also. What I wanted to showcase was that the same option is available for our Oracle database as well. And for our for a lot of our newer databases, such as SAP HANA or MongoDB, Postgres, all of that, we do provide the option to perform UI based recovery, making sure you have the same benefits as SQLs and of the world. So again, like I just went to my SAP HANA instance or other system. And then once I click into it, I will be presented with the same calendar view like you saw for the SQL host. Now, again, if you click in, you can see that we again have log backups enabled. So you can slide the basically the timeline view to select the most recent or the timeline from where you want to actually perform the recovery form and then click on recover. Now for SAP HANA, it's one of the newer workloads that we have started supporting. So you will see less options for recovery. So we do have the option to restore to original database, which is more like in place recovery. Or if you want to export the database, you can, you know, choose the option to select to a different database. Now, the same options are again available for our MongoDB's and even for our Postgres SQL host as well, Like I mentioned before. So, you know, the idea here is that Rubrik is trying to unify a your protection of databases, different databases. And the same thing is true for your recovery option also. You know, whether you want to perform recovery for your HANA, Mongo or for your SQL databases, the look feel is the same and we do provide you with granular COV options so that you don't lose data while recovery as well. And with that, I will hand it back to you, Mick. All right. Well, thanks Alpica. Appreciate you walking us through that. Now where I wanted to go next is, it's all well and good due to recovery. You know, recoveries are important and obviously the the call it the tactical execution of the recovery is what we're always focused on. But I think there's always a broader question of once we get to the recovery, like when you start planning out a recovery or figuring out, we want to answer a few questions that we know strategically what we're trying to recover. And usually this revolves around if there's an attack or if there's a disaster, have to ask ourselves, okay, what was affected database wise? And how do we prioritize amongst those databases? Because even though, obviously, with a lot of tools and a lot of advantages with Rubrik, we can accelerate the recovery, The trick is is there's still work that has to be done. We still probably have to organize those recoveries, prioritize which databases we're bringing up first. And then finally, in the event of a disaster, there's always gonna be questions after the fact of what was affected, what was impacted, particularly when we start thinking about ransomware and the potential for exfiltrated data. So auditors are always gonna have questions about, can you highlight what database data might have been exposed in the process? This is where Rubrik introduces our sensitive data discovery. So we've actually had this in the platform for quite some time when it comes to unstructured data. So files, text on a system. Basically able to scan this, match it against certain discovery patterns, identify certain markers, and we've done this across the entire platforms to where we can scan your systems, looking for anything that might be a ransomware deployable, working off of your backup history, giving you the flexibility to create your own scans as well as using existing ones. So this has been a cornerstone of Rubrik for several years now, but in the last year, what we've been able to do is introduce this functionality, particularly around sensitive data discovery, to the database data. Now, when we look at this feature, what Rubrik has done is we've leveraged that sensitive data engine that we use for the unstructured data. Again, something we've had for a couple years now, And then integrated that with our database protection systems. So what we'll do is you can go into Rubrik and we'll see this as part of the demo where I can open up the interface and go in our sensitive data monitoring. And basically, I can now configure for my databases broadly, hey, what policies do I wanna scan against? You know, use either existing ones for PII, PCI, HIPAA, you know, canned information like that, or we can create our own, right? We have the ability to custom create analyzers and rules for doing the data classification. Once we have those rules in place, we then can apply those to the database so that when Rubrik does a backup of your database, we'll then follow that with a scan against the sensitive data in your database. Now what's cool about this is we're not actually we're gonna scan the data in the database. We have to, right, to match against those patterns. But, you know, obviously, there's always a concern of, well, I don't want to expose that data. I mean, the whole purpose of it being sensitive is we wanna try and secure it. Well, Rubrik doesn't actually store the data itself. It will scan that data, but it will only track the metadata of that. And we use that those analyzers to categorize the data, to classify what's out there, and basically store just the results of that scan. Once we have those results, we can then pass it to our engine and say, okay, does this violate a policy? Does it not violate a policy? And we can very easily then display for you the results of your scan so that you can obviously see where your sensitive data lies. So with that, you know, just a few minutes to kinda touch on that really magic since seeing what it actually is. So I'm gonna pass it back to Alpica and she'll walk you through the sensitive data discovery functionality. Thank you so much, Mike. Let me quickly share my screen again. We are back to our data protection dashboard, which is part of Rubik's Security Cloud. And sensitive data discovery is actually part of our Data Security Posture application. So, we click the app tray and head to data security posture. What you see here is, you know, a lot of graphs and how we get here. I'll talk about in just a little bit. I just want to set some context before we get here. So for that, let's head to policies and data classification. Now, Rubrik actually has a data classification engine. That consists of a lot of analyzers. These analyzers actually map to different sensitive data types. So, Italy's driver's license, US social security number, passport number, bank routing information, these kind of things. Now, they are very geographic oriented as well. So, for EMEA, you know, we might have a different type of analyzers versus for the rest of the world for the APG portion, we might have different analyzers. What you see here is predefined. We have hundreds of them. Now, what if you have a different data type sensitive data type that's not covered by Ruplit, right? Very simple. You just simply click on create data type and give it a name, select a data category. It could be data category is essentially the policies that perhaps do. I'll get to that in just a little bit. We'll leave this as it is. Select the sensitivity. So if it's medium, high, low, non sensitive, those kind of things, let me select sensitive. And then you can either use regular expressions or dictionary words to actually define your new data type or analyzer. So supply the value and you can even validate it before you actually create it. It's as simple as that. So once this is done, you we had two data categories. Now what are data categories? These are the policies that Mike was talking about like HIPAA, your CCPA financial, these kind of things. Again, like you do have the option to have predefined Rubric provides you with some of them. Now, if you don't think this is enough and you have a special category to which your data maps to Fairly simple. Create category. Apply all the analyzers that are either predefined or it could be custom. Select one or many of them. And I'll just select two. Select what the document types are. That's optional. And then just give it a name. You can then assign this policy to your workload. So very simple process like financial. I'll just take that. And then assign resources. So we're talking about databases. I'm going to select my database right here. I can select the object or I can apply to all of them. I'm not going to do that right now. And that's it. So every time a backup happens, Rubrik will go ahead and also look for sensitive data for that particular database. Right? Now, once that is done, the results of the classification is what is displayed on this dashboard. So like you see here, there is SQL and then there are other workloads as well. We will focus solely on the database aspect of it. So for that, I'm going to just quickly go object inventory and filter by SQL. Okay. So once that's done, what you see here is that that there are a lot of databases where there is personal information or geographical data present and rubric has identified it. Let's take an example of this cork object. Essentially, it's the instance, the SQL instance, and it can have multiple databases, right? So it lists the instance and all the databases where this sensitive data is present and it will list exactly which data category or the policies the sensitive data adheres to. So on the side, like essentially on the right side, you can see for the DBO database, what are the data types? The data category is personal. So there's last name as well as EU email address, right? So it needs to be protected and making sure that only the right DBAs have access to it. So, with this, you can use this information to lock down your databases and make sure that only the right applications and right people have access to it. Now, what can you do with this information? A couple of things. You can download a CSV file from here. You can give it to your DBAs or your legal team or your compliance team to make sure that these databases are completely locked down and only the right people have access to it. And if there is a cyber attack, it can help you identifying sensitive data exposure as well. So, you know, before you have a cyber attack to limit the amount of sensitive data exposure, you can use this information to lock down your databases as well, like I mentioned before. Now, the other thing is that currently sensitive data monitoring is supported for QUIL and Oracle databases only. And our plan is to come out for other databases as well. But for now, it's for SQL and Oracle. And with that, I will hand it back to you, Mike. All right. Well, thanks Alpica. Honestly, I'm really excited about the Sensitive Data Discovery. I believe it unlocks a lot of capabilities for not just DBAs, but for other folks within the IT enterprise, right? People who may not have access to the data directly but still need to recognize or report on where sensitive data lies. So this is a really cool feature that really can unlock a lot of functionality around your data and really give you more insights than you may have today. It's all well and good for Alpeak and I to sit here and go on about the greatness of Rubrik, which, you know, I think it's pretty great. But it's always good to see it validated out there in the landscape around us, right? So these are just some basic numbers that we have pulled from our customer base, right? Most of our customers, and we're several thousand customers out there using Rubrik globally. And the vast majority of them are doing some level of database protection within Rubrik. Within that protection, we're protecting somewhere around two and a half million databases, again, globally, various sizes. And what's cool about the last two numbers, right, is you can see we scale up and we scale out, right? So the largest database we protect with the financial services customer is over a petabyte and a half. That's a single database. That's not a group of databases. That's not a server. That's a single database. But then, in addition, like I said, we scale out. We have one customer that has almost 100,000 databases in their environment that they're managing Rubrik. So, like I said, you can see the validation here. It's one thing for us to walk you through the tech, show you the value that we see, but it's another for us to show you, like, we had a lot of customers out there that are validating it on their own in their day to day use. Always do like to touch, so there might be asking here. We've made mention of a couple different database platforms. So as of today, these are the database platforms that Rubrik is currently supporting. We support Oracle, SQL Server, Postgres, DB2, HANA, as well as Azure, Amazon RDS, MySQL, and of course MongoDB for NoSQL databases. Now we are planning more. Keep your eyes open, keep your eyes peeled on Rubrik because we're always looking for other databases that we can add to our support. But right now, this is what you get with Rubrik for today. And of course, coming back to the validation, these are all the customers that are actually protecting databases with Rubrik today. So customers like KeyBank, like Home Depot, Ulta Health and Beauty. Basically, all verticals, we see a lot of use of Rubrik for protecting databases, recovering those databases, and basically using Rubrik to ensure the continuity of their business. With that, we'll go ahead and close out for today. I wanna thank Alpica for doing those demos and thank you for your attention as we go through this. Feel free to reach out to anyone from the Rubrik team if you have further questions or you wanna explore more. And of course, we have some things on our own website that you can browse into. With that, we'll go ahead and close it down. Thanks again and have a great rest of your day.