The Economics & Beyond of Object Storage

Are you getting pressure from finance to justify or reduce your cloud bill? Cloud Zero is the only cloud cost platform loved by engineers and trusted by finance. Cloud Zero can identify unused, idle, or over-revision resources, alerts you to spend anomalies and organize 100% of your spend into a framework that mirrors your business structure, like cost for customer, product feature, or team. It's the most powerful platform ever built to provide accurate, granular visibility in your total cloud spend without the typical pitfalls of legacy cloud cost management tools, like endless tagging or clunky Kubernetes support. Managed cost, optimized development, and maximized profit, all in one platform. Join companies like Rapid7, Drift, and SeekEek by visiting cloudzero.com/cloudcast to get started. That's cloudzero.com/cloudcast. Visit today to experience immediate and ongoing savings on your cloud bill. Cloudcast Media presents From the Massive Studios in Raleigh, North Carolina. This is The Cloudcast with our endelb and Brian Grayceley, bringing you the best of cloud computing from around the world. Good morning, good evening, we're back to the cloudcast. We're coming to you live from our massive cloudcast studios here in Raleigh, North Carolina. And it is Aaron for Cloud News this week and it is the beginning of June. I can't even believe that. But we've got some really good cloud news articles this week, but I also feel like we need to mention the elephant in the room. Certainly not Enterprise Tech, but certainly Emerging Tech Apple announced their AR VR headset. I don't know. Won't be out for another year. When it does finally come out, no one will be able to afford it. So we'll see. I don't know what else to say beyond that. So we're just going to jump right into our news this week. First article this week, Reddit announced earlier that they are going to start charging for API requests. And this is a trend we're starting to see. And this is a larger shift in the market. A lot of organizations that were supported by ad revenue previously, some of that ad revenue with the changing economy starting to dry up a little bit. So what are some ways that they can make money? Well, API is kind of makes everything go round. We've talked about it many times on this show. It's, it's, you know, the way organizations are talking to each other. It's the way everyone's accessing their data. And oh, by the way, it's where a lot of the security impacts are happening as well. But Reddit is looking to charge somewhere in the neighborhood of $12,000 for 50 million API requests now. That, you know, sounds 50 million. That sounds like a big number. But, you know, for instance, there was a vendor in the article that they would be looking at $7 billion requests a month. So about $1.7 million per month. They're about 20 million per year for them to continue to do API requests against Reddit. So some pretty significant impacts to this. And there's lots of protest think about it. There's lots of like subreddits going quiet. And it'll be interesting to see where this ends up. Now, the next one. And this is actually, I don't know who is in charge of PR over at core weave, but kudos. So on Wednesday, they announced $200 million dollar funding round. And if you're not familiar with them, think of them as a vendor that specializes in providing and partnering with Nvidia for GPUs. And really powering a lot of the folks that are looking to do open AI and chat GPT in particular. So they announced a $200 million funding round on a valuation of $2 billion. And then they also announced, hey, by the way, we're partnering with Microsoft. And so Microsoft actually probably secured this earlier this year, but it came out as well at the same time. Microsoft is struggling with the demand of open AI. So how can they ensure pipeline and ensure continued growth? Well, they partnered with core weave. So really good, interesting partnerships happening and you know, funding rounds happening. And core weave, for instance, you know, brought out the new cutting edge and video GPUs. And I talked about that in a cloud news article gosh, I don't know two, three months ago, somewhere thereabouts. So good for them. They're certainly killing it. And they had a hell of a week. And for our third and final article, GitLab, GitLab really did well from an earnings standpoint. They reported earnings and they beat guidance. So what do you always want to do when you're reporting earnings? You want to beat and then you want to raise expectations. Well, they did both. They beat guidance and they raised expectations for the year. And because of that, their stock went on a healthy jump and certainly good for them and wish them continued success. If you want to dig into the financials deep, you can of course look at the cloud news, show notes and this and all the other articles are in depth in there. So with that, I'm going to wrap up cloud news for this weekend. Coming up after the break, we're going to be talking about economics and beyond about object storage. And we're going to be talking to John over at Cloud Ian about it. Today's episode of the cloudcast is sponsored by DataDog, a real-time monitoring platform that unifies metrics, traces and logs into one tightly integrated platform. DataDog APM empowers developer teams to identify anomalies, resolve issues and improve application performance. Begin collecting stack traces, visualizing them as flame graphs, organizing them into profile types such as CPU, IO and more. Teams can search for specific profiles, correlate them with distributed traces, and identify slow or underperforming code for analysis and optimization. Plus, with DataDog APM live search, you can perform searches across the full stream of ingested traces generated by your application over the last 15 minutes. Right, DataDog APM free with a 14 day trial and DataDog will send you a free t-shirt. Visit datadog.com/apm-cloudcast to get started. That's datadog.com/apm-cloudcast. And we're back. And what we're going to do this week is really explore a topic that's as a technology bit around for quite some time. But we want to do a little bit of a different spin this week. We want to talk about object storage, but we want to talk a little bit about economies of around it and what has kind of happened over the last 10 plus years. And what the future is for object storage as well. So in order to do that, we went out and found an expert like we tend to always do here on the show. And so we have John Tor, CMO at Cloudy in this week. Welcome to the show, John, and tell everyone a little bit about your background. Well, hey, thank you very much. It is great to be here. I guess my background could be really summarized very quickly. It's basically all storage all the time. I began my career actually as a design engineer on hard drives with C-gate. When C-gate was a tiny little company based on a small town near here where I lived in Las Gatos, they were in Scotts Valley. And my storage career progressed through hard drives in this engineering side into hard drives on the marketing side and then systems on the marketing side. And that's really kind of where I've stuck ever since. I've done a few different startups. I've been involved in a few large companies. So a lot of variety, but it kind of revolves around storage. Fantastic. And Cloudy is really interesting to me for multiple reasons. And I believe like your career in Cloudy and kind of parallel in many ways. Cloudy, first of all, has kind of been around, I'll say, since before object storage was cool. When it was kind of an emerging technology way back when, and we first heard about Cloudy and back in the open stack days, and for longtime listeners, they'll remember that. And early AWS S3 days. And so, object storage certainly has come a long way. And so can you help everyone frame like, where were we and where are we today? One thing that's generally true about storage technologies is they tend to have a very long gestation period. They sit on the shelf being relatively unused for quite a long time. Before a use case comes along, it really drives adoption. Fiber channel was exactly the same way. It's sat essentially on a shelf out in Minneapolis for many, many years before parallel scusy ran out of gas and then became a need to do something new and different. Well, object storage is the same way. Object storage, as you rightly pointed out, started a long time ago, it actually started as a way of making data immutable. It was introduced by EMC commercially as a system to protect healthcare records so they couldn't be changed. And that was, you know, definitely a use case, but it wasn't going to drive any large scale adoption. Well, what happened was as we progressed through block storage, file storage, and then into the cloud, the cloud providers recognized that they really needed a extremely scalable storage environment to meet the needs of their cloud customers. You were not going to be able to build a hyper scale environment on file. In fact, the last major cloud provider that was built on file was Yahoo. Every major cloud provider after that was built on object storage. So they saw the need to do things at a totally different scale. That drove adoption of object storage. And what came along with it was that object storage became endowed with a whole bunch of cloud capabilities. So you got really two things. You got the underlying technology object storage and then you got this protocol, this API, which was now adopted as Amazon AWS S3. And that endowed object storage with a whole bunch of cloud capabilities that made it multi-site, multi-tenant, gave it multi-part upload capabilities, which was really important for moving large files over the internet. So you got really this set of use cases of cloud demands that made object storage something completely different than what it started out as and really made it the storage of the next generation. - And I'll add to this too. Like yeah, I think the object storage to your point was always there, but I think it was the S3 compatible APIs that really solidified the use cases to kind of across the board, right? And so ever since then, we've seen this rise of what I'll just term enterprise class S3 compatible object storage, right? And that's where the use cases really started to explode. You see hybrid cloud GC, especially data sovereignty these days with data sovereignty laws. And most recently analytics, such as data warehouses, data lake houses, kinds of things popping up as well. And so where are you seeing implementations these days and 'cause we've really moved beyond basic simple storage for the cloud back ends, right? - Absolutely. And that's been a really interesting evolution around this. And storage is always driven by the use cases. You have to have the use case demanding capabilities of this class of storage, before our customer's gonna go out and buy it, Mary. No one adopts a new technology and then waits for the use cases. They encounter the use cases and then they adopt the new technology. So that's exactly what's happened here. And what we've seen is that as these use cases have become cloud enabled, that's really what's driven the adoption of object storage. And let me tell you what I mean by that cloud enabled. So backup is a classic example because when you're looking to back up your data center, you need a target, right? That data has to go somewhere and cloud emerged as a good target for backup data. So as a result, all of the enterprise backup vendors adopted the S3 API as one of their supported protocols so they could use the cloud as a backup target. Well, obviously once you've adopted the S3 API as a protocol, you can now put that data anywhere that has S3 compatible storage. It doesn't have to be in the cloud, it can be on-premises easily. So because the backup vendors all went and did that, it emerged backup emerged as one of the key use cases for S3 compatible storage on-prem. Well, we've seen that same thing happen, as you already said, with data lake houses, analytics, all the vendors in that space also are looking to make their software be cloud enabled. And so that's another emerging use case. And we're seeing this across the board, we're seeing this in M&E, we're seeing this in healthcare, we're seeing it in all kinds of data analysis software. Everyone is looking to make their software cloud enabled, but when they do that, they're also making it just S3 enabled so we can deploy S3 compatible storage on-prem and get this exact same benefit. >> And kind of aside to that, 'cause you did mention this earlier as well, and just kind of popped into my head. Tell everyone a little bit too, because I think again, go back to the early days of all of this, and it's the more simple storage aspects of all of this. We started to see some security concerns early on, and then you start to see data privacy, the rise of that, and as I mentioned data sovereignty as well. And so how big of an issue these days, oh, they're then obviously very large, is security in all of this, and how has the evolution of security as more copies of data kind of get in more places? >> Yeah, it's huge. The security issues and concerns that customers have are huge, and interestingly enough, this is an area where S3 compatible storage really can shine, because, first of all, think about ransomware protection, for example, to protect your data from ransomware, the best way to do it is to make it immutable. If you can make a coffee of your data that's unchangeable, ransomware cannot go in and encrypt your data. That's not possible. And so as a result, your data is protected from the most common form of ransomware hacking. Well, as I said up front, actually that was the original use case of object storage was immutability. That's what EMC sold it as 20 plus years ago. And interestingly enough, AWS, with S3, improved on that with a standard for how you manage that immutability. They created a protocol, object lock, that made it possible to manage that immutability feature in a standardized way. So as a result, the backup vendors adapted object lock as one of their APIs. We implemented object lock, and it's been a key driver of data protection on S3 compatible storage. Because you can protect your data from hacker intrusion, it solved one of the big problems we saw happening when COVID started coming along. When COVID started, we saw an enormous uptick and ransomware attacks because of people working from home. They became more vulnerable to email phishing attacks, et cetera. So that became a huge driver. But security overall, and this kind of gets into that next thing, you make how we're as object storage going. Object storage is becoming the de facto for private clouds of all kinds. And data sovereignty is a major use case for private clouds. So we think about governments or highly regulated industries who need to keep their data in specific environments because of regulatory concerns, just fear of pure geographical boundary concerns, they want to create a private cloud to manage that data. Obviously, what they're concerned about it with that data is sensitivity, right? The government's got government records, it's got health care records. So protecting that data becomes job one. You've got to make sure that data is secure, and it's got to be bulletproof. So you want to be secure, but at the same time you want it to be accessible because you're putting it in this private cloud, the whole point is to make it accessible by the people that need to use it. So you can't rely just on the confines of the data center. You've got to have security protocols that are best in breed and can be used by when they're accessing the data from anywhere. Well, that's exactly what we have implemented with Cloudion. We have more security certifications than anybody else in the business. These are government security certifications set by the SEC, set by the Department of Justice, set by the government agencies that buy this kind of storage and use it in their private cloud. Cloudion was the storage of Mill Cloud 2, and we just to announce we want to deal with the Department of Defense for their new version of their government cloud. So security had been part and parcel of our message for years now, and we worked very hard to enhance that story. So basically what we've got is a private cloud, which is as secure or more, then the best, most secure data environments you can buy anywhere. And what that means is now you can set up the sovereign cloud environment in your data center and know that data is protected from ransomware attack. It's protected from external attacks, from people coming in from the outside. And it's protected even from people within your own data center. We locked down the box itself by disabling root access. So even if you have physical access to the box, you don't have a way of tampering with it. So we've addressed this from all levels. And what that's done is it's made us a great solution for these sovereign cloud environments, and we're currently in environments in Germany, in New Zealand, in Australia, in several different companies in APAC, countries in APAC, other places in Amia. I mean, that sovereign cloud has emerged as being a really important use case for S3 compatible storage. And security is one of the key reasons why. - Fantastic. And I'll add this too, like we'll dig into another topic here, which is cost and cost effectiveness because If I look at, you know, like take the last Qubicon here in the United States back in Detroit, one of the big, you know, themes, if you will, was cost. Another big one was security. And especially with the economic environments changing slightly, like interest rates starting to go up and, and you know, VC Monday, VC funding isn't quite as loose as it used to be. And interest rates going up, you know, there's a lot of folks really starting to figure out, okay, how do I secure everything? But how do I also make sure the economics makes sense? And so we're starting to see this more in design conversations, but I'll just say there's always a healthy tension if I use that term between what's technically possible and what's economically feasible just because something can be built doesn't mean it's economically viable at times. And so how does that kind of design conversation play out lately? Well, that's been a very interesting dynamic over the past, you know, three, four years. Object storage has always been a highly cost effective platform because it's built on industry standard hardware. And that's been true, you know, for many, many years now. So our underlying hardware platform is very cost effective. Our software itself tends to be very efficient and low overhead in the way that we, you know, store the data and protect the data. So the cost of object storage versus other storage types has always been very, very competitive. In fact, we're substantially less expensive than than most file or block systems. So if you just compare apples to apples, you know, our storage class versus others. That's, that's always been an easy conversation. You're going to, you're going to save, you're going to save money with object storage. The conversation that's become, you know, right in front of center of people's minds has been, well, what about cloud, right? Because if I go to cloud, then I'm purely on a op-x model. I only pay for what I use. And, you know, I can have the, you know, flexibility of growing without ever having to lay out any additional capital expense. And that's, you know, that's obviously been a, you know, interesting conversation. And at the start of COVID, that conversation became, it took out a new dimension, which was my on-prem data center has become hard to manage, right? Getting people to go into my data center to run it, to install new stuff to maintain it has become difficult. And so that made the cloud that tip the hand for the cloud during that conversation. Fast forward a few years though. Now we're, we're kind of moving out of that era. What we've seen is this. What we've seen now is that, you know, on-prem object storage and cloud have both become so similar in so many ways. We use the same API. We offer the same kind of enterprise scalability. We really, We really made a very different kind of storage environment that parallels very much capabilities of cloud. But with one key difference. And that is, you know, we are primarily a, a capex model where you're buying the equipment up front, whereas cloud continues to be an op-ex. So depending on your use case, one could end up being substantially less expensive than the other. And what we're seeing as customers are now evaluating their use cases, they're now looking at, hey, if I'm, if I'm, if I know what I'm going to be consuming, I know I can predict my storage requirements. Because I'm managing a large data set that is of a fairly known size. It's likely it's going to be a lot less expensive for me to operate my own environment than it is to go out and buy it essentially in a rental, you know, buy it by the drink in a, in a public cloud environment. So we're seeing now as customers are making that assessment rather than just saying, you know what I'm having trouble running my own data center because of staffing because of, you know, COVID constraints. Now they're saying, I need to make a decision per use case on where it makes sense for me to run this workload. And that decision could be driven by data sovereignty, a decision could be driven by cost, a decision could be driven by latency, right, depending on where that data is being sourced and where it's being used. And based on those requirements, I'm going to make a decision on where I should run that environment. So it's no longer just a, you know, let's just let's just go to the cloud. Now I've got a choice of exactly where I run this and I don't have to make a choice based on technology because I can run the same data types. I can run the same applications on Prem or in the cloud. And I can manage them the same way. So technology is no longer a, you know, a necessarily a driver in this maybe in some cases, but in many cases, I can do the same thing either place. So it really comes down to where does it make business sense to run this application to run this workload and, you know, what's the best way of going and doing that. And that's where object storage is really kind of emerging as being the on Prem cloud. And that's why hybrid cloud. It's such a key part of our messaging because what we're saying is if you got workloads that can run, you know, in a cloud, enabled way. You have a choice. You can run those workloads on Prem. Do the math. Is that that makes sense for you or you can run them in the cloud. And what we're seeing now is customers are very much making that choice based on some, you know, pretty serious analysis of their workloads. And I'll add maybe a potentially another vector to that because this is what you wouldn't be, you know, as you were talking about that. There was a term that was really, you know, thrown around a number of years ago and I haven't heard it nearly as much recently and that's the term of data gravity. And so the concept for those unfamiliar with it is this idea that data has a certain amount of weight, a certain gravity. And it attracts more data to it, you know, existing sources become hard to move over time. And then more and more tends to get added to it. And and so what we, especially with I would say the rise of hybrid and multi cloud. Folks don't think about it or talk about it as much and we used to talk about, oh, make sure you're placing this in the right locations and you had had, you know, the ingress in egress charges and all these other, you know, things to think about. Well, now we're in an environment where we are seeing more hybrid and multi cloud solutions. We are seeing larger and larger data sets. And so where are folks building these large data sets? And where, what are they building them for and last question, you know, are they ever moving them or is it, you know, more of the latency is good enough kind of scenario. Yeah, data gravity is definitely a driver, you know, large, large data sets are difficult to move. And, you know, is a result of, you know, the industry had developed all kinds of physical work around putting putting storage and trucks and moving them that way, you know, the old, the old joke, you know, you can't beat the bandwidth of a, you know, a station wagon full of tapes moving down the highway. That goes back a lot of years to when we had station wagons. But the same is true now, you know, even even the post station wagon era, people are still moving their data and trucks for the same reason. And they, and they have, it does happen and data gravity is absolutely a concern when people are collecting a lot of data in one place. You know, it is certainly easier to just leave it there and operate on the data there. And now because you have the ability of running that workload, whatever that workload is, you know, if it's a cloud enabled workload, you have the ability to run that date that workload, wherever the data is. That decision just got a whole lot easier, right? You no longer have to worry about how am I going to move that data? You say, well, I just need, I can bring a workload to the data. I don't have to change the word. I don't have to change the software. I can run things the same way. I can just do it close to the data. So that conversation is actually gotten easier. But it's also taken on new dimensions because data gravity is not, and now it's not only the, you know, the size of the data, but it's also, you know, the physical location of the data versus, you know, in relation to what that data is. So TikTok is the classic example right now, right? All this user data. And the question now is where does all that user data reside? And so that's what TikTok's, you know, project Texas is all about is, we're saying, well, that date that user data is now going to reside in Texas, because that's going to alleviate privacy concerns. And that's really become more the compelling issue today about where, you know, where does that data physically reside? And how do I make sure that I'm complying with, you know, the data sovereignty requirements. So I would say data gravity, you know, ability to move data is probably important data sovereignty today is probably even more important. Latency is always a concern, right? If I'm if I'm running a workload in a clinic, for example, I'm trying to access health care records or images. If those images are stored in the cloud, how long is it going to take? Well, Well, if that cloud is in, for example, in Singapore, where the AWS region is for that area, and my application is in Vietnam, and I'm having to move that data back and forth in undersea cables, what's the latency involved in that? How quickly am I going to be able to access that data? So that's definitely a point of conversation today as well. And then finally, the cost. So I guess the point is that data gravity now has kind of multiple dimensions. And all of these things are saying, you know, I need to factor all these things in when I decide where I'm going to run the application. And that's where hybrid gets so important because now you can say, well, I can run that application wherever it needs to be. I don't have to change my process. I can run the application close to the data and the conversation just got a whole lot easier. And I'm potentially saving myself some money at the same time. So it's a win-win. Very nice. Very nice. So, John, one last question for you. And maybe this is also some advice for other folks starting companies or thinking about business. We like to approach the business side of this as well. And so, Claudian is really well known for their partnerships, their alliances and their solutions. You have quite a mix. And part of that maybe just goes back to being a startup that has been around a lot longer than other startups. But if you look at the mix, right, there's hardware companies, there's software companies, there's backup companies, there's public clouds. I mean, there's a little mix of everything and anything in there. And so, has this been, first of all, intentional and a factor in Claudian's longevity? And tell everyone a little bit about how this came to be and how important you see this for the future. Yeah, we are so very proud of our Alliance partners because they're really key to our business. And really, I can think about these in three categories. The first would be their Alliance partners that are related to our use cases. So the most obvious example would be backup where we work very closely with Veeam and Rubric and Colin Boughton Baratoss. Working with those partners is so important because that are our ability to integrate with them to provide a solution which is plug and play and really satisfies the customer's needs. And in the least cost, in the most cost effective and the most performant way has been really critical. So working closely with those guys has been so important to us. And then we're seeing it now with the analytics vendors, working with folks like Teradata, Enverdica, and Cribble, and Splunk, and working in the analytics world. Microsoft SQL is now a partner of ours. Microsoft SQL adopting the S3 API. There's an amazing one for you. So working with those partners has been a key to driving our use cases. The second thing is channels because we need routes to market and we are always sold as part of a larger environment. Obviously, it's never a standalone thing. It's always a larger solution. So working with folks like Lenovo and HP Green Lake, we're now part of the HP Green Lake environment has been so important because there are great channels into that market. And then the third area is kind of a catch all and that's hybrid cloud. Because as the cloud providers start looking at how they provide services closer to their users. So they can address data sovereignty concerns, for example. We are a solution for that S3 storage that is local to the user. We're working with AWS on outposts, which is their hybrid cloud solution that's on prem. We're also working with AWS on local zones, which is their sort of mini region that sits closer to the user, but provides capabilities just as if you were in an AWS region. So we are storage for outposts. We are storage for local zones. So working in these hybrid cloud environments with AWS and others is a really critical one. And I would put in that same category, we work with cloud providers worldwide to provide cloud services that they go, they turn around and offer, including some of the largest in the world, like IONOS over in Amia, massive cloud provider that relies on cloud and has to underline storage infrastructure for object storage. So those hybrid cloud environments, private cloud environments are the third class of major partnership. All of these are really critical to driving our business and we work really hard to make sure we have a well integrated and well support and solution for each of them. Fantastic. Well, with that, I think that's a great place to really close this out. John, I just wanted to say thank you very much for your time this week and your insights on all of this. And everyone that's out there, thank you very much for listening and if you enjoy the podcast, please leave us a review wherever you get your podcast, please tell a friend and we certainly want to continue to grow the community. And if you have any feedback, certainly reach out at any time. And so on behalf of Brian and myself, thank you everyone for listening and we'll talk to everyone next week. Thank you for listening to the cloudcast. Please visit thecloudcast.net to find more shows, show notes, videos and everything on social media.