But what if your data is too big for the Cloud?

Cloud visionaries promised us unlimited scalability, but they greatly underestimated the amount of data we would start producing.

Today we sit down with Michael Tso, the co-founder of Cloudian, and he explains why some systems are just too big for the cloud, and how the industry is adapting. Mike also shares his advice for selling via channel partners, and we talk about the competitive advantages and disadvantages of being a Japanese startup on the global stage.

It's a great conversation, and I think you'll enjoy it.

Show Notes

The real reason we are "drowning in data" The missing link in connecting cloud and local storage What kind of apps really need 100's of petabytes of storage How to pivot successfully in Japan, and why it's so hard to do so Using Japanese culture as a competitive advantage The most important difference between Japan and US startup culture Why US companies hesitate to buy from Japanese software companies How to expand globally using channel partners and systems integrators What everyone gets wrong about failure

Links from the Founder

Everything you ever wanted to know about Cloudian Friend Mike on Facebook Follow him on Twitter @MichaelTso

Transcript Welcome to Disrupting Japan, straight talk from Japan’s most successful entrepreneurs. I’m Tim Romero and thanks for joining me. As a society, as a species, we have way too much data on our hands. A decade ago, our data got too big for our local systems and so we moved it into the cloud. And now, well, our data has gotten too big for the cloud and we're moving it back on premises. Now, I promise this will make a lot more sense in a few minutes, when we sit down with Michael Tso, co-founder of Cloudian. Cloudian makes massive scale storage systems, massive as an hundreds of petabytes of storage that can run on site and seamlessly integrate with Cloud Storage. Cloudian is also interesting, because they're one of only a handful of Japanese startups that have really succeeded in the US and European markets. And Mike and I talk a lot about how they made that happen. We talk about the challenges and the necessity of pivoting in a Japanese startup, how Cloudian's Japanese identity and culture both helped them and hurt them in their global expansion. And Mike gives some really great advice about how to sell software through channel partners and what you should really expect from those relationships. But you know, Mike tells that story much better than I can. So let's get right to the interview. Interview Tim: I'm sitting here with Michael Tso, the co-founder of Cloudian, which provides object storage for the enterprise. And Mike, thanks for sitting down with us. Mike Tso: Yes, happy to be here, Tim. Tim: Now, object storage for the enterprise is something that gets talked about a lot. But what does that mean exactly? Mike: Yeah. So I think I need to just rewind a little bit and just kind of talk about why are we drowning in data? Why is there so much data? Well, because actually, if you think about the process of innovation and the process of invention, it's data-based, right? You need to have data to know where you're at, and you can draw inspiration from analyzing your data. And then you go try some invention, and then you have to measure the outcome. And then you go back and kind of either fix it or improve it. So it's a continuous process. Tim: But, I mean, if you look at kind of like human history, we humans will fill up as much data, well, we've always had data, it's a question of how much we can store. And the easier it is to store, the more of this information we will. So it used to have to be engraving on cuneiform, clay tablets and creating vellum now it's just, oh, we'll automatically just save everything. Mike: Well, I think about this a lot, right? It starts with how easy it is to create the data. So if you have to chisel away on a rock, you're not going to create a lot of data. But these days, we have sensors with cameras, when people think about data, people think about, oh, how many emails am I sending? How many pictures are you taking? But actually all these are very small in the grand scheme of things. What is generating what the big data that we're seeing these days are mostly machines, because we have automated data curation, we've automated data movement, we've automated data storage. So, the whole thing is becoming very easy. And now we have AI, which has just a unsatisfiable appetite for more data, because the more data you have, the better your AI becomes. Tim: We've got this perfect circle. So we've got machines generating the data, the machines storing the data and the machines analyzing the data. Mike: That's right. And thankfully, humans are still needed in a process to make sure everything works. But exactly, the more value you're able to get out of your data, the more people are keeping all the data, so we're not keeping it just because it doesn't have value. We're keeping it because people are really generating this data. So actually, in a week from now, I'm going to be speaking at a conference called the High-Level Forum at Grenoble, France, where the theme is reinventing industry. And of course, I am there to speak about what I think is at the center of all this, what's underneath all the reinvention is the amount of data that people's collecting so that they can invent better. So anyway, so coming back to the object storage, when you have that much data to manage, you really have two or three kinds of sets of problems. So data that's growing about 20-40% a year, that's growing 5 times to 10 times faster than the IT staff that we have in the world. So you need a solution that can manage sort of an infinite amount of data with the same team, right? And that's one of the things that object storage allows you to do. And then secondly, something like 80% of all new applications are born in the cloud. So, all these application expect cloud-native environments, and that includes object storage, because every cloud is using object storage as its way of storing data. And third is security. We're seeing 700% increase in attacks from ransomware. Building strong firewalls are really no longer sufficient. You have attackers that only needed see once, and once they come in, you have so much data that they're going to be kind of going after it. So, the data storage, the data management systems need to be self-protecting. You cannot rely on building bigger walls. Tim: To frame what Cloudian does. So, Amazon, Google, Microsoft, they're all talking about object storage. Mike: 90% of data on Amazon, Google, Microsoft is stored in object storage. Tim: What does Cloudian add to this? Mike: We develop object storage software that allows you to have the same kind of technology that's being used in the cloud, but you're able to deploy this in your own data center or in somewhere else where your data cannot move into the cloud, either for cost reasons, or performance reasons, or privacy reasons, or legal reasons. So really, the way to think about what we do is that we provide an extension of the cloud and bring the cloud to where your data is. We bring all the flexibility, all the ease of management, all the stuff that the cloud offers, but we bring it to where you happen to have the data. Tim: Okay. So, if I'm a programmer sitting at one of these companies, and I'm accessing data, it all looks like I'm just accessing data in the cloud. I don't necessarily know if it's on-site or off-site or, okay. Mike: Exactly. The idea is that your data should be resident wherever it's most efficient for you. Our software makes it looks like it's a cloud-native environment. Tim: Awesome. You know, Mike, we've known each other for a long time, and I've been trying to get you on the show for a while, because you and Cloudian are one of the few real success stories of a Japanese startup that's succeeding in the US market. I want to get into that, but first, I'm still trying to wrap my head around the scale of this data. So, you guys are talking about applications in the tens of petabytes. Mike: It's actually hundreds now. Tim: Or hundreds of petabytes. Mike: Yeah. Tim: And so when I was doing research, trying to understand what you do in this industry, I'm trying to wrap my head around these numbers. And it just gets silly, you know, you're talking about, well, it's millions of songs or hundreds of thousands of videos. The Library of Congress is about 20 petabytes of information, and like, my brain just can't wrap around that. So really, what kinds of things require tens of petabytes of storage? Mike: I can tell you quite a lot of examples. So one of our customer that has hundreds of petabytes of storage, they're a major media company, and we archive for them all of the media that they have ever produced from the beginning of time. People tend to think about movies or TV shows. They think about, okay, it's going to fit on a DVD, right? Okay, it's true, that's the finished product. But actually, all the clips that it took to make that, all the takes that they have, they actually keep all of it, because a lot of times they go back and they do a remake or they do a recap, or they make the movie for a bigger screen or for lower resolution for mobile devices. So there's lots of different versions of this. And with 4k here and 8k coming, you're going back and going to the source files, and they are making higher resolution versions of the same movies or the same shows. So, before they put us in, they used to keep that library on a bunch of tapes, and they have to tape libraries—one that they use actively, one that they live in like a mountain that they never touch. When you're using tape, you know, when I want something I think it's on that tape, tape number whatever, I send like a truck to go and get it.