Revolutionizing Cloud Storage: Backblaze's Impact on Data Solutions and Enterprise Growth

Unlock the secrets of cloud storage evolution with insights from Pat at Backblaze, as we journey from their humble beginnings in 2007 to becoming a force in the industry. Discover how Backblaze's commitment to affordable and secure data solutions has not only transformed individual backup services but also enabled large enterprises to thrive. We'll unravel the significant impact of their S3-compatible API launched in 2020, which revolutionized seamless integration and skyrocketed adoption rates. This episode sheds light on Backblaze's proactive approach to safeguarding data against threats like ransomware while maintaining transparent pricing and unparalleled interoperability, making them a standout choice for businesses navigating the complexities of modern cloud storage.

Explore compelling use cases from businesses utilizing Backblaze B2 to propel their data-centric applications forward. From optimizing security camera systems to supporting AI development, learn how diverse industries manage vast data volumes with remarkable efficiency. We'll tackle the challenges of international expansion, like data sovereignty, and highlight strategic moves such as opening a new data center in Toronto. Whether you're a large enterprise or a startup considering a cloud storage migration, discover how Backblaze B2 offers a smooth transition and significant cost savings compared to other providers. This episode provides essential insights for businesses eager to adapt to the fast-paced digital landscape with cost-effective, modern cloud storage solutions.

Transcript

Speaker 1 0:03

Hey everyone, we're diving into the world of cloud storage today, beyond just backup and what's needed to take storage to the next level. With Backblaze Pat, how are you?

Speaker 2 0:16

I'm well, Evan. Thank you very much.

Speaker 1 0:18

Well, thanks for being here. Let's dive right in. Maybe talk about Backblaze and your core mission and solution, and how do you fit in this very busy and crowded landscape of cloud storage these days?

Speaker 2 0:35

Right? Well, Backblaze was founded 2007, so quite a while ago now. So quite a while ago now. And the original mission was for ordinary people to be able to back up their Macs and PCs, and we had five co-founders that bootstrapped the company from this idea of affordable backup. You know, having a I think originally it was $5 per month for unlimited data, because most people you know, we know how many gigabytes we have stored on our laptops Most people don't know the difference between a gigabyte and a megabyte or how many they have. So their idea was like let's make this a service and you pay five bucks a month and we back up your data. You pay five bucks a month and we back up your data.

Speaker 2 1:25

And in 2016, they introduced Backblaze B2, our cloud object storage offering, and in 2020, introduced the S3 compatible API on that. And that's really where adoption lifted off, because all of those off-the-shelf tools and SDKs that use Amazon S3 work with Backblaze B2. So we IPO'd, I think, just over three years ago, towards the end of 2021. And it's pretty incredible Between founding and IPO, I think they took about $5 million of funding, which is, I mean, I would love $5 million, but apparently it's not a lot for that process and where we are now.

Speaker 2 2:24

We've got over half a million customers, over three exabytes of customer data under management, and last quarter we signed two seven-figure deals with customers. So from charging individuals five bucks a month, we're now up to working with large enterprises spending in the millions of dollars per year, or over two or three years.

Speaker 1 2:58

Brilliant. Well, let's talk about the large enterprise market, where there's a lot of opportunity and challenges around data protection and security, and you know, everyone who understands ransomware and how it works and cyber threats knows that. You know securing your data protection is fundamental to avoiding those catastrophic data breaches and losses that we read about in the paper every day. So how do you see yourself fitting there? And you know how do you add value to those enterprise clients who are facing these threats.

Speaker 2 3:32

Sure. So there's two sides to this. There's the technology and the functionalities that supports that, and then there's how Backblaze is different. And there's a contrast there there, because on the functionality side we're very much the same. It's the same apis, uh, like the object lock functionality where you can say, um, on a file by file basis, you know, this file cannot be deleted until this date, and there's literally no instruction you can tell us, there's no support call you can open, that will delete that data and that integrates with the corresponding functionality in backup tools. So, veeam, you'll see it called immutability, where you can back up your server and say, okay, this cannot be deleted for three months and it can be extended after that three months, but you can't turn it off. And really that's the key to that.

Speaker 2 4:49

Um, you know, cyber security, ransomware protection, because even if some piece of malware were to somehow, um, you know, infiltrate your system, obtain the keys that you use to uh backup data to backblaze b2, it can't uh delete thoseutable backups, that object lock. There's just no instruction you can send to us that will cause us to delete that data. So you know, and that's pretty industry standard, you know, the S3 APIs become like a de facto cloud storage standard and then looking at what makes Backblaze different. We are very cost-effective for those backups. You know it's $6 per terabyte per month. It's a fraction of the cost of Amazon S3 and the other hyperscalers, and particularly for this backup data, that's a very compelling offering for a lot of enterprises where they realize just how much money they're spending with AWS just how much money they're spending with AWS.

Speaker 1 6:11

Yeah, cloud sprawl and storage sprawl is a real challenge, and so it looks like you're known for your transparent pricing, and I guess you interwork or interoperate with all the hyperscalers, whether it's Azure, aws, google Cloud these days it really doesn't matter for you and your customers.

Speaker 2 6:35

Yeah, I mean. The reality is that all of the off-the-shelf and custom products all tend to use the Amazon SDKs and we are API compatible. So whether you use those SDKs, whether you write directly to the endpoints, we just work the same as Amazon S3. It's not a complete clone. There are little corners where there are different functionality, but for the 99% case, you change, you know, you create a Backblaze account, a bucket, an application key, and you configure your. You know the product that you're using and the key thing is you need that custom endpoint to say, okay, you're working with Backblaze B2.com rather than Amazon AWScom. And then, as far as from you, far as from the tool you're using whether it's a backup product or a media asset management system or a custom app it just looks exactly the same.

Speaker 1 7:39

Brilliant. Well, that's good for customers, for sure. Speaking of which, certain industries have really compelling needs right now for backup and recovery. I'm thinking healthcare in particular. But where are you seeing most growth, most traction, most need across different industries these days? Most?

Exploring Cloud Storage Use Cases

Speaker 2 8:01

need across different industries these days. Yeah, I mean, it's so. Obviously, backup and archive has always been like the backbone of the business. But very quickly we found that there's a much wider use case for cloud object storage. So we have a lot of customers working in media and entertainment where it's not just an archive for like footage, it's actually their working storage. So they're uploading footage as it's being captured into Backblaze B2, and then their tools whether it's a media asset manager, the Adobe tools are accessing Backblaze B2. So it's their working storage that they're collaborating around. And then we have what we call application storage customers where their actual core business revolves around storage. Core business revolves around storage.

Speaker 2 9:11

So a great example is we have a customer that does security cameras for commercial premises and one of the options you have with their product is upload footage to the cloud. So we have I think now in the order of tens of thousands of cameras sending footage directly into Backblaze B2. So it doesn't bounce through. You know our customer systems. The cameras are sending data directly to Backblaze and then you know if their customer one morning the alarm went off at 10.30 last night. I want to review the footage. You know that Backblaze customers portal lets them log in and then it's calling our APIs to bring back the bit of video footage, maybe download it as an MPEG. And you know that's a growing area for us Customers who actually have data-centric applications that are based around Backblaze, and that's a trend we're seeing now as we've seen the growth of AI over the past couple of years.

Speaker 2 10:21

We know we have customers with you know, throughout the whole AI lifecycle, from training to inference, log data. You know every conceivable point, checkpoints, every conceivable point in the AI data lifecycle. Our customers are using Backblaze and sometimes we're surprised by you know what we're seeing here. You know I wrote up an internal document explaining. You know, and this was some time ago, this was over a year ago. You know how customers are using Backblaze for AI and the number of times I had to go back and revise things as we talked to more customers and they're like well actually, you know, you have this training process where it's really interesting.

Speaker 2 11:18

Actually, I went to GTC, the NVIDIA conference, last year and heard this training process described. And these are big systems with many GPUs and it's actually tough to keep them running for any length of time. There's always something going wrong. You've got so many components in the system, so these systems regularly take checkpoints. So training a model can be a weeks-long exercise and if it fails two weeks into the four-week training, you really don't want to go back to the start. So they'll take a checkpoint every hour, every half hour or something that they can pick up, and we saw customers uploading the checkpoints into Backblaze. It makes complete sense. But from looking at it from our point of view, how is this going to work? It's just something we hadn't considered. Our customers are constantly surprising us with how they're using our platform.

Speaker 1 12:22

I can imagine and you work internationally as well where you get into lots of interesting and hairy situations with data sovereignty and other regulations.

Speaker 2 12:33

I imagine you have to help navigate all of those global issues as well, and US West regions, so several data centers in US West, one in US East. We have our EU Central in Amsterdam and just I'm not sure if we've officially announced it we announced that we were doing it, but we've opened our Canada East region in Toronto and that was in response to Canadian customers or Canadian prospects saying I would love to use you guys, but I can't because it's south of the water. Yeah, so we opened our Toronto data center, like literally over the holidays.

Speaker 1 13:24

Brilliant. So any advice for business or technical leaders looking to kind of get started with a more modern cloud storage strategy. I mean, how quick can you move? What are some of the pitfalls or roadblocks that might need to be overcome and move into this new era?

Speaker 2 13:44

Well, I would say well, first of all it depends on how much data you have. It depends on the use case, on you know. On the one extreme, big company with you know a lot of moving parts probably best giving us a call. We've got a great sales engineering team. Our system engineers are very experienced in looking at cloud architectures and figuring out, okay, how can this work better than it does right now. But then on the other end of the spectrum we've had self-service customers that have run you know their own POC and just migrated to.

Speaker 2 14:33

We've got a really interesting customer, Amplify Data they're called. They do a data distribution platform and it's built around Snowflake. That's their data hub and their product is SAS. Providers can incorporate that data hub kind of porting to their site so that you can give analysts access to data in a number of different formats. It does like clever things like zero copy joins across different snowflake accounts. But the 80-20 case for data analysts is can I download a CSV file? And they were on Amazon S3. They had their startup credits and when they burnt through those they discovered that, wow, this storage stuff is quite expensive. And we discovered this all after the fact.

Speaker 2 15:36

I was looking through our logs and I noticed some data coming from Snowflake. I wonder what's happening there. And I got in touch with literally the holder of the email on the account and sent an email saying I'm really interested in how you're using Snowflake. You probably used the article I wrote on how to do it, but it turned out that they literally just found us through a Google search and got it working just by creating an account, a bucket, application key, setting up their software to point to us instead of S3. And the CTO said it was two weeks from having the idea to being in production to being in production. So you know, that's one end of the spectrum where it's a startup with a well-defined product that has like one route into the storage.

Speaker 2 16:36

You know Snowflake exporting files into Backblaze V2. But yeah, in general the data migration is very straightforward. The devil's in the details around that it's when you start talking about okay, well, where's the data coming from? Is there a cost to get the data from? You know my hyperscaler, for example, into Backblaze B2? And then is there a better place I could be generating that data than, say, aws, where I've got to pay $0.09 per gigabyte to get?

Speaker 1 17:20

it out Still charging, still overcharging after all these years.

Speaker 2 17:23

Sorry, my friend at AWS.

Speaker 1 17:26

We won't go too deep into that, I'm sure.

Speaker 2 17:29

Yeah, so you know sometimes it's a simple okay, we just move the storage and change the endpoint URL and it just carries on working.

Speaker 2 17:40

And then sometimes it's well, actually let's move the compute to, say, vulture, one of our compute partners, where we have a bandwidth agreement so there's no data transfer charges. Maybe let's move our CDN to Fastly and do a re-architecture that ends up being something quite different and sometimes an order of magnitude more cost-effective. And you know we call that approach the open cloud, where you know you're not just locked into a hyperscaler for everything that you can pick and choose technologies from providers. So maybe your storage from Backblaze, your CDN from Fastly, uh, gpus, from, uh, say, core weave, and you know we've got those partnerships, um and I. You know part of what I do we didn't talk about what, who I am and what I do but part of what I do as chief technical evangelist is, um, writing those recipes and reference architectures and code samples to show, okay, this is how you get things hooked up and run your compute here and your storage here and your CDN, and it all comes together into a solution.

Speaker 1 19:01

Wonderful Best. Of Bree is still alive and well and I assume I'm a small fish, but I assume I can just get my credit card out myself and sign up and stop being oh, no credit card. Well, stop, certainly. Apple asks for my credit card for storage every three months. I need to upgrade, or?

Speaker 2 19:20

Google Cloud. You can get started with an email address and we give you 10 gigabytes to play with for free, and that first 10 gigabytes is always free. And then you know, when you want to say, make a bucket public, that's the point at which we say, okay, we need to know a bit more about you. So, you know, take a dollar off your credit card just to kind of like, make that link, um, and obviously when you go above that 10 gigabytes, uh, you know that's when, uh, you're into the pay as you go, um, and then we have, you know, like, uh, variable costs don't work for everybody. So, as well as the pay as you go, we have a capacity-based pricing called B2 Reserve, and that tends to be when we're in those enterprise deals with the sales team and it's not just an individual putting their credit card down. We're talking about sometimes petabytes of data, of course.

Speaker 1 20:25

Well, one day I'll be at the petabyte level. Until then, thanks for joining. Really intriguing, I'm here at CES and probably a big consumer demand. Still, I know you're focused on the enterprise, but lots of opportunities as we seem to be consuming and creating infinite amounts of content these days. Thanks so much, pat, for joining. Really appreciate your insights and chatting.

Speaker 2 20:48

Yeah, my absolute pleasure and yeah, I hope to see you again sometime.

Speaker 1 20:53

And thanks everyone for listening, watching and sharing. Take care. Cheers.