#050 – Data Protection and Kubernetes Resilience with Michael Cade & Julia Furst Morgado (Veeam)

Itiel Shwartz: Hello everyone and welcome to another episodes of Kubernetes for humans podcast. Today in the show I have two guests and not only one. so if you can please introduce yourself. Julia you start.

Julia Furst Morgado:

Michael Cade: Yeah. Hi thank you for having me. It’s a pleasure. My name is Julia and I work with Michael at Veeam and uh my I we’re probably going to talk about my journey but I come from a non-traditional background. I used to be a lawyer and then I worked in marketing for a few years and then got into this world of cloud-native you know Kubernetes platform engineering DevOps. uh at first was very overwhelmed but um but yeah I’m very involved in the CNCF community. I’m a CNCF Ambassador. I’m also an AWS Container Hero. I work with with also like Civo Ambassador ambassador for several programs. I really like not not just working with technology but I think I like working with the people behind the technology. I think that’s why I think it’s so important to be in these communities going to events. But yeah, uh I can talk a little bit more about how I got started later on and after Michael introduces himself.

Itiel Shwartz: Yeah, Michael, go for it.

Michael Cade: Yeah. Hey everyone. Yeah, so I come more from a CIS admin type background, virtualization, storage, data center type workloads, backup. I’ve been a Veeam for just over 10 10 years now. So live and breathe backup and data resilience. But the inflection point around platform engineering, CIS admin or is just is a CIS admin ultimately just evolving into that that that platform engineering group, keeping the lights on, getting stuff out there, automating everything, orchestrate everything. And and that’s really been my last five years at Veeam is trying to uh evangelize and and advocate for different ways of thinking about how we how we consider data protection data when it comes to platform engineering. Normally when you’re speaking to platform engineering team they don’t necessarily think too much about the data and the storage.

Itiel Shwartz: Oh no no data is another team like a bit different team. uh but but before we’ll jump into like the technical of data I will be happy super happy to hear about like really like the journey of both of you so if like each can do like in two minutes so m I’ll let you start this time and if you can like in two minutes like like it feels like a bit less tradition more traditional right like adment talking about platform engineering is a bit less unique but where did you start what happened and how come you are in Veeam like for 10 years so if you can like share on top of the

Michael Cade: Yeah. So, so 16 years ago, 17 years ago, I started off on the help desk helping both end users and businesses with their computer problems. Right? These are people in different parts of the business that don’t know what they’re doing. and then I moved into more focused areas around storage and virtualization, VMware and NetApp storage. And that’s when I came across Veeam. Veeam was interesting to me because it integrated into everything. It was the Switzerland to any platform, any any storage and it had a lot of integration, bells and whistles to to make that happen. So I was interested in getting into that. I think more recently it’s been more around like M&A and being involved in the acquisition of Kasten which is focused on protecting Kubernetes and then that leaprogged me into the world of DevOps creating a a a little project called 90 days of DevOps um just as we came out of of uh the pandemic. But yeah, that that’s me in in two minutes as to where how did I get here?

Julia Furst Morgado: Yeah. And for me, I would say it was kind of a serendipity how do you say serendipity? I was like I said you know meeting people in the community I was doing a coding.

Itiel Shwartz: No, no, no. You have to start before that. Like you said, you’re a lawyer. I know.

Julia Furst Morgado: Yes. A lawyer anymore.

Itiel Shwartz: That’s not going to be two minutes.

Julia Furst Morgado: But yeah, trying to

Itiel Shwartz: highlight highlight

Julia Furst Morgado: highlights. Yes. I was a lawyer in Brazil. I’m from Brazil originally. And when I moved to the US, I studied business and then I started working in marketing. And my last job was at an MSP, so managed service provider. And I knew I wanted to become more technical. I just didn’t know how how to but I saw there was a coding boot camp online that was go about to start for free and I jumped on it and I started you know studying by myself uh coding mostly with a boot camp but also learning cloud very little Kubernetes. I didn’t know much about it, but because I was, you know, networking with people online on social media, Twitter was a thing back then. This was four years ago. now it’s it’s totally different, but I was meeting a lot of people online and I met someone online and they came to New York, you know, for work and we ended up meeting in person and they gave me a ticket to go to KubeCon. So that’s how that’s how I got started. I wasn’t planning on, you know, getting my hands dirty with with cloud-native tools or anything. I was planning to become a software engineer. But then going to KubeCon, it was, you know, enlightening. I saw there is a whole new world there and so many people passionate about these technologies and I decided to to jump on it and start learning studying uh you know hosting my website on Kubernetes and you know uh little little game server and small things obviously always breaking things and trying to debug a YAML but I think that’s that’s my journey and also contributing a lot to open source. I speak four languages and I’ve been doing a lot of translation from documentation also uh white papers from uh the TAG uh working like working groups and TAGs um so helping with that and I think connecting people in the community people that need help with something you know mentoring other people in the community so I think that’s how I got started and uh now I work with Michael we’re on the same team at beam and you know um working with with a great team around Kasten as well so backup for Kubernetes and trying to raise awareness about the need for backup because it’s the last thing people think about especially when we’re talking about you know cloud-native applications etc uh most of them are stateless and now they’re starting to run databases on Kubernetes etc. But we go to these conferences KubeCon uh KCDs to to talk about that because it’s really important. Yeah.

Itiel Shwartz: Okay. So like a very interesting story. I feel both of you like talked a bit about Veeam but maybe like give some more like context of you know I think for like the average person it’s not necessarily like a household name. It’s in my head somewhere there next to NetApp and Nutanix. I might get it like completely wrong and I if you don’t mind like explain how old is the company, what are you guys doing? Why are you doing that? How did Kubernetes change stuff? So, let’s do a bit on the company itself. Yeah, go ahead, Michael.

Michael Cade:

Michael Cade: Yeah. Yeah. So, so Veeam’s Veeam’s just coming up to or around 17 years old and Veeam started off protecting virtual machines specifically on VMware and then Microsoft Hyper-V and it was the virtual machines and obviously we all know the applications and the data the databases that lived on top of those machines and how critical that infrastructure or that wave of virtualization was. So Veeam first started in in regards to protecting those in an application consistent way being able to recover them as fast as possible. Then that fast forward 17 years we’ve got multiple platforms multiple hypervisors. I’ve been calling it the hypervisor hunger games with what’s going on with um the current virtualization landscape including VMs on Kubernetes as well as VMs in the public cloud. equally we’ve got cloud-based services, public cloud-based services, RDS, Azure, SQL, Google, blah blah blah. so they need protecting SaaS-based workloads, Microsoft 365, um, Salesforce, Kubernetes, like they’re all different platforms that run our workloads on a day-to-day basis. I think it was easier when I was a CIS admin 20 years ago where I just looked after physical physical data centers, physical servers, maybe some virtualization and I had to protect that. Now we’ve got data all over the place. We’ve got data in SaaS PaaS like you name it, you’ve probably got some data somewhere. So where Veeam have started off as virtualization backup, it it’s it’s extended out to being able to protect workloads, hitting the line of um security. so you’re not backing up bad data, you’re not recovering bad data. But interesting that you mentioned like the NetApp and the Nutanix. So NetApp to me, they’re a storage company. We’re agnostic. We’re not storage. You bring your own storage. Put your put your backups somewhere that you want. That could be object storage. That could be a DAS device, a NAS, a SAN, etc. but we’re not storage. You’re going to bring your own storage. You’re going to choose. Nutanix, however, obviously a bit more of a platform play. They have their own hypervisor that we support. They equally have their own virtualization and Kubernetes um distribution as well that that we support from a Kasten. But it’s about all of those are just different platforms that we that we protect. So, not as not as like for like, but interesting that you’ve you said that because I think that’s a that’s an interesting take.

Itiel Shwartz: I don’t know like you say data to me like that that’s like the first like that’s the first things that come to mind.

Michael Cade: That’s where you saw it, right?

Itiel Shwartz: Yeah. So, so maybe like share a bit like why why do I need you guys or like in general like not even you like you know I’ll give you like a real example like right I work in corridor. I’m the CTO here. We don’t use you guys like we don’t use anything other than like AWS. You should. Yeah. So what am I what am I doing wrong or what should I start doing like starting tomorrow?

Michael Cade: So I think the first fundamental reason why you need a why you need to consider resilience of data. It’s not just about backup and recovery. That’s when bad things happen. But equally it becomes part of the security posture as well. Equally around portability. But I think the one thing that I’d say to you is like what happens when bad things happen and again if I go back to CIS admin days bad things happen was misconfiguration accidental deletion natural disasters fire flood and blood. Now in 2025 we have the the very topical uh concept of ransomware or cyber security.

Itiel Shwartz: People are are attacking supply chain. they’re they’re attacking all over all of that data, they’re trying to exfiltrate that data. So backups become even more important because if something’s encrypted in your production environment, then you want to be able to bring that back. You brought and that’s more general across all platforms. You mentioned AWS and you’re probably using AWS Backup. Now with AWS Backup, you’re and I’ve not got anything against AWS. They’re very good friends of ours as they are everyone’s. But one thing from an from an AWS Backup

Itiel Shwartz: AWS is like controlling like all of the market. But yeah, we’re all here like good friends of AWS.

Michael Cade: Yeah. So, so from an AWS Backup point of view, they’re going to be storing it in EBS snapshots. Like so snapshots are great for really fast recovery, but they cost you a fortune. I’d hate to think or maybe I’d like to think how much uh AWS actually make on on snapshots. So where Veeam come in we offload that off to to cheaper deeper storage into AWS S3 or even even better off to a different location completely so that you can recover that workload wherever you need it need it to be and then that brings me round to the last point around portability portability. So if you’ve got EC2 instances running AWS um AWS Backup funnily enough is only going to let you recover those workloads into AWS. Veeam gives you the ability to recover them into other public clouds, other hypervisors even in a way we can get it onto onto Kubernetes as well.

Itiel Shwartz: Okay. Okay. Now I think I’m starting to get it a bit more. so I have like a couple of question. There’s like an Israeli company named Eon or something like that that does similar stuff or I’m like completely again like uh not not not part

Michael Cade: I think I’ve heard E, right? Yeah, like those guys sold the backup company to AWS and like they’re like back then they talked quite a lot about backup, but uh maybe it’s not really what they do. no, I’m looking them up. It says turn cloud backups into large strategic assets. I think yeah, that’s something they they do backup as well, but no, I don’t know them

Itiel Shwartz: well. Yeah, like it wasn’t like a pre question again trying to to better understand. So what you’re saying just to to take what what you said right now, maybe I’m doing stuff correctly, right? Like I do have quite a lot of backup for pretty much everything in Commodore. What you’re saying is that you are able to save me a bunch of money and maybe make it a bit more portable, which for me is not that big as I’m quite I am quite hooked to AWS already. So like backup won’t be like the last thing that will stop me uh from not using AWS.

Michael Cade: yeah and and that’s fair enough, right? I think it’s all about choice. the other the other area that we see a lot is um customers they get from on-prem into the public cloud into AWS and they’re using EC2 instances and then they’re putting their database on top of that. AWS Backup has no inside information about that VM and the database. So you could take a snapshot as much as you like and this might lead us into the Kubernetes world of data services as well, but

Itiel Shwartz: AWS Backup is just going to look at the image based layer. It’s not going to be able to quiesce the workload. It’s not going to be able to talk to Postgres or Oracle or blah blah blah. so it’s just going to be a like pulling the power cable out of the back of the server.

Michael Cade:

Michael Cade: And that’s going to be the point in time that you get. Whereas what we do is a little bit more application consistent. Dive into the application, tell the application this is what we’re going to do and then caress that workload and and take that take that back up. But if you’ve already moved on to your like I would always say that EC2 is might be the start of the game in in the cloud, but you should always be looking how do I how do I get those databases out of an EC2 and into RDS or into Dynamo or into something else.

Julia Furst Morgado: And even if you’re running them on RDS or Dynamo, we also protect those. We protect even VPCs that a lot of uh companies they they don’t protect, but you know it’s helpful in case of you know a disaster. You can like quickly get back get your configurations back. But uh it’s good because you have everything in one central place you know uh you can manage everything from like uh the Veeam Backup & Replication and uh especially for for a backup ad administrators it’s much easier than having to go to AWS to to take care of AWS Backups and then they have workload somewhere else and they they need to manage another tool. So with VI it’s centralized in one place you know you we have also um like uh using security tools to to see if scanning all the backups and everything now we’re implementing AI obviously everyone is so uh I think everything all all of this is is uh helpful uh to to the data the backup teams backup administrators

Itiel Shwartz: Okay, that makes I think like total sense now. Maybe like like maybe like one or like one one more thing that I’m trying to understand. You said that like uh it should like interest more like platform engineers or or something like that. Maybe if you can share like who should care about that like I’m the CTO, right? Like I I do sign quite a lot of documents. So disaster recovery is quite important for us and that is like one of the primary use cases that we do you know we do backups but who is like the typical buyer of something like that who should care about that if I’m opening a startup how soon should I have backups and if you can like shed a bit more color on like how does this product actually fits in the normal like dayto-day.

Michael Cade: Yeah, and it’s a good question because I think we see we see a bit of everyone. Like it’s not it’s not every not every CEO or CTO even cares about data either. Like they will do when something bad happens because it’s their business that that’s going to go down or going to have problems. Generally speaking, if there’s not a backup team and again with a startup, you probably don’t have a dedicated backup team.

Itiel Shwartz: I don’t have a backup team. No, no, I don’t

Michael Cade: backup person even then like it’s it’s it’s the role and responsibility of whoever is is in charge of that data and that might be the application owner. So what I’m getting at is is really it’s it’s anyone anyone that’s responsible for data should have that mindset of being able to protect it. and we and we work with it used to be we’d just walk in and speak to the backup administrator or backup team and that or the virtualization admin or the cloud engineering team. Actually, the the world of siloing things like DevOps platform engineering, SRRES, systems administrators, it’s a much easier talk track because application DBAs. It it really depends who who cares. I would always say that we’re bottom of everyone’s Christmas list. Backup is boring. and I say that with a with a smile. I’ve been here 10 years. but equally, it’s the it’s the insurance policy that we all have to have. Like, all of us have to have insurance on our cars, regardless of where we are in the world. We have to have it on our house. Like, it’s the same for backup and our and our data. it’s just whether someone’s had an experience that means they need to protect that workload or know that they need to protect that workload. So, I I guess maybe given that you’re you’re conscious of backups and data protection and disaster recovery, then maybe something’s happened in the past that that gives you that that inside information.

Itiel Shwartz: I’m not sure if it it really happened as much as like a working companies that were very like backup um like u aware as well as I feel a lot of it come from regulation also a bit like insurance like in Israel I’m not sure again how is it in the states

Itiel Shwartz: but in Israel you have to have an insurance right like you buy a car you have to like buy an insurance if you do SOC 2 to disaster recovery I’m quite sure that I signed a bunch of stuff regarding like my backup, right? Like I’m I’m quite sure that there were I have like a whole document on what happens when things go goes out. So like I wonder like do most of your customer come from we have a compliance to meet or it’s after the fact I don’t know like after they delete all of their like production database and they forgot that they don’t really have a backup. So yeah, like I

Michael Cade: I think Yeah, I think it’s a bit of bit of everything really. Like people people only put people only do things after they’ve been bitten, right? They don’t do it to

Itiel Shwartz: to to be react reactive and proactive. I’m telling you like like SOC 2 is a really good reason to do a bunch of stuff that are good like SOC 2 has a lot of downsides I think but it does forces you to to confront some of the real like issues right like DR is again it’s it’s important right like not only from a backup perspective from the process and how do you do it and how do you set up but yeah sorry

Michael Cade: so I so I think yeah I completely agree and I think even more so across EMEA um more recently ly more like the last three to four years we’ve seen a lot more regulation directives around data and how we use data where we store data and I think that is a driving force to a lot of the a lot of the features and functionality that we’ve built into our our platform but equally it’s a it’s a forcing function for businesses to to to do something rather than just leave it at the bottom of the shopping list.

Itiel Shwartz: So, so maybe share a bit like on the you know like you are there for 10 years and Julie like feel free to to jump in like but but what changed like if we look at the last 10 years right since you joined the company what are the latest trends what do we see where are we going to be maybe like in a couple of years in this like backup space like

Michael Cade: yeah like what’s interesting

Michael Cade: so I think so when I started it was just VMware it was VMware it was Microsoft Hyper-V but it was generally VMware every every single shop that we spoke to would have VMware and virtual machines. Then I would say probably starting I would say 5 years ago we started to see the adoption or at least that that start of Kubernetes becoming a uh in in those in those companies as well and the public cloud probably probably eight years um SaaS as well. So not only so what I’m saying is so virtualization kind of was the start for me at Veeam but ultimately then it became multiple platforms. We have workloads here there and everywhere but there’s still the same underpinning requirement of that insurance policy still exists. we just have to talk to these different platforms potentially in a different way. like one of the key differentiators that Veeam came out with was when when Veeam first started protecting virtual machines on VMware, we didn’t install an agent like our our competitors at the time. We just went into the vSphere API and we did it agentless and and brought in things like change tracking and everything that we did that way. very similar to what Kasten did from a Kubernetes perspective is Kubernetes gets deployed within your Kubernetes cluster becomes part of the Kubernetes API via CRDs and aggregated APIs and you can obviously interact with a nice flashy web UI um or you can interact with it via the Kubernetes API using kubectl or whatever tool you want to to interact right and and what I mean by that is that where we’ve gone after these mult multiple platforms to protect them. We’ve just done that in a methodical way, a more efficient way to what needs to happen versus, oh, we’ve got an agent that backs up these physical machines. Let’s just throw that agent on a virtual machine. There’s obviously a lot of ways in which you can you can move data. It’s really it’s how you recover that data when bad things happen. So being part of that API or using that API is the is the win from that perspective. I think I think what we’ll see uh like just very quickly in in what we’re going to see over the next 18 months two years is more adoption around Kubernetes probably that the hype around AI workloads and data in intensive workloads is only going to drive the adoption there. I think we’ll also see some more um SaaS SaaS-based applications that that we that we offload a lot of that admin overhead off to. we’ve already seen it like Salesforce and Workday. I’m sure the Microsoft 365 we could you could name Riverside. We’re literally using Riverside now whereas however long ago we would have been using OBS or or something else, right? So

Itiel Shwartz: also VMs on Kubernetes is a a big thing as well with KubeVirt. Now we’ve been seeing a lot, right? Yeah. Yeah.

Itiel Shwartz: So So Julia, maybe like a question for you because I know that you mentioned that you love speaking. How do you take something like backups which again like I don’t feel it’s like necessarily like the sexiest thing and like how how do you add some interest into it if like if any like so so maybe

Julia Furst Morgado: I add AI to it. Oh, I’m joking. But uh yeah, usually it’s talking about data protection, the need of it, and usually I show headlines from newspapers about ransomware attacks that happened in in that certain location where I’m speaking because it happens a lot every day, more than we think. And sometimes you know um e companies they get attacked over and over because uh you know they they uh recover their data but the data is infected. So I think you I don’t talk much about Veeam when I’m giving talks at KubeCon or or other events but mostly the need for you know data protection backup uh you need to think about that uh it’s usually the last thing companies think about I was talking to a friend he’s a a senior platform engineer at a a very big company enterprise and he said, “Oh, I want to I want to hear what Kasten do, you know, because we have everything set up, but we don’t have backup yet.” So, I was like, “Oh my god, you know, it’s so important.” You’re an enterprise, so yeah, I think it’s mostly about raising awareness about the need of data protection.

Itiel Shwartz: No, that that makes total sense. So we don’t have a lot of time but uh you know like you did mention like a bit the AI boom right so maybe like what’s what is really going to change maybe on that regard so if let’s say that all I want to do is like a Kubernetes and inference and training like that’s it so now do I need backup more or less like it feels to me a bit more emperor so maybe the data is not that important but Maybe I got it wrong. So

Michael Cade: yeah, I think there’s there’s two aspects to that in my opinion. And one is how much did the vectorizing or the inference cost you? if that process is very quick then and it didn’t cost you a load of GPU and CPU turns and cost you a lot of money then probably you don’t need to protect it. It’s more ephemeral at that point. Right. However, if that did cost you a lot of time, effort, money in resources, in time etc. then it probably should be backed up. The way in which I look at AI infrastructure is twofold. One is the vector database is just a database, right? It’s full of data that probably needs protecting if it’s important data or it can’t be recreated quick enough. I think the other area of of what I what I’m seeing is how do we protect the whole configuration of that pipeline? It isn’t it isn’t just a a an infrastructure as code or a Terraform state file that looks after this. It’s there’s generally a lot of configuration, a lot of complexity that builds up these AI pipelines, the building out training models, etc. So there’s twofold. Do you protect the database? That’s I don’t I can’t answer that. it depends. And the second thing is the configuration of that. And again, that depends. Hopefully, you’ve got that in some sort of source code repository and you just play it that way. But they’re the just the conversation. I honestly don’t care whether you buy Veeam or not. I don’t get paid on it. So, it’s it’s it’s about just raising awareness of when you should back up stuff.

Itiel Shwartz: Okay. I think like that’s pretty much it. Julia, less remarks by you.

Julia Furst Morgado: I mean, um, I just think it’s, uh, important to to think about that and I’m glad that we were able to to chat and and talk a little bit about Veeam, what we do there. And, um, we’ll we’ll be at KubeCon. So, if anyone wants to chat with us or connect with us on LinkedIn, uh, or talk about anything else besides backup, we talk too much about backup sometimes. So, yeah. anything else community open source. I’m I like that those as well.

Itiel Shwartz: Okay, that’s super cool. So with that, I think we’ll finish and thanks a lot.

Julia Furst Morgado: Thank you. Woo!

[Music] Kubernetes for Humans.

This is an AI generated transcript of the conversation