By Ben Golub
Over the course of my career, I’ve had the unique privilege of working for six startups, three as CEO. While there are no guarantees in the startup world, there are few things as exciting and powerful as a talented team working on a compelling and disruptive solution that solves big, important problems at a time when multiple trends depend on those problems being solved.
Storj Labs meets all of these criteria and more, which is why I am excited to say that I will be joining their incredible team and community as Executive Chairman and interim CEO.
Why Storj Labs
It’s clear to many that we are on the cusp of a major new shift in computing, driven by a fully decentralized internet and by new, decentralized models of trust and security, powered by blockchains and distributed ledgers. When I first talked with Storj, I was intrigued by the team, and by the fact that they had already built a robust platform and passionate community to provide the storage layer for this new, decentralized internet, as well as developers that wanted to use this new layer. Put briefly, Storj Labs is a decentralized cloud storage provider. Rather than own its own data center, like centralized cloud storage platforms, Storj Labs works with individuals and organizations (called “farmers”) who rent their excess hard drive capacity. This capacity is utilized and paid for by users in need of object storage. For a broad and growing set of use cases, the resulting network is more secure, reliable, economical and private than traditional centralized cloud storage platforms. As the network scales, it generates compelling economics for users, for farmers, for partners, and for Storj Labs itself. In conjunction with partners, we are making decentralized storage part of a broader decentralized stack for distributed applications by acting as a base layer storage infrastructure. While Storj Labs is still in the early stages, it is already providing secure, reliable, performant storage for over 60 petabytes of data from 10s of thousands of users on a network of 150,000 decentralized storage nodes run by 10s of thousands of farmers in over 200 countries and territories worldwide—all without operating a single data center. In many ways, this looks like my early days with Docker. At Docker, we used to talk about building a programming layer for the internet, allowing people to build applications and services that could run and interoperate seamlessly across the millions of devices connected by the Internet. The impact Docker made was largely driven by building a huge community and turning that into a compelling business model.
Storj is similar in that it is building a storage layer for the internet that will leverage the huge scale, diversity, and hyperlocality that the millions of devices connected to the internet provide. And, we are doing so by embracing a broad community of developers, of open source companies, of customers, and of farmers, bringing uniquely compelling economics to all and building a great business model for ourselves in the process.
Why is this important?
The world is creating exponentially more data, most of it in the form of files and folders. Gartner estimates that, by 2020, the world will create over 44 zettabytes of data a year. To put that number in perspective, if you tried to store 44 ZB on CD-ROMs (which are 1.2 millimeters thick), you would need a stack that reached to Mars and most of the way back. True, some of that data is cat videos. But, that 44 ZB also comprises the data that is key to curing diseases, to piloting self driving cars, to keeping planes in the air, to finding new energy sources, and to building new and compelling businesses.
Gartner also estimates that over half of that data can’t and won’t be stored in traditional data centers and centralized cloud services.
There are many reasons why. First, economics. While compute prices (i.e. what it takes to create data) reliably halve every 18 months due to Moore’s law, storage prices dropped only 10 percent in the last 18 months. Some of this has to do with industry structure, but a lot has to do with the fact that electricity and people don’t follow Moore’s law. Second, the centralized cloud and data center model is ill-suited, from a security, performance, and reliability basis, to many decentralized and distributed use cases.
The decentralized cloud storage that Storj provides has powerful benefits:
Reliability - Done correctly, distributed, broadly decentralized storage is more reliable and resilient than traditional, centralized storage. In a representative use case, an individual file on the Storj network is encrypted, sharded, erasure coded, and distributed over 84 nodes. Over half of those nodes would have to fail for the file to be lost - something that is extremely unlikely to occur when none of the nodes are correlated by using the same location, infrastructure, shared power, etc.).
Security – The Storj model is also more secure, by design. There is no central location to attack, only the user knows where each of their shards are stored, and all the data is encrypted. To compromise a single file, a would-be hacker would have to locate multiple shards spread across nodes in over 150,000 unique locations in over 200 countries and territories. They would have to compromise all those nodes to retrieve the encrypted shards. Then, they would have to decrypt the shards. And they would have to repeat this process for each new file. As one security expert stated, “Storj is like spreading encrypted sand on an encrypted beach.” As the network grows, it also becomes more difficult for attackers to locate individual file shards.*
Performance – For many workloads, especially for large files, Storj is vastly more performant than centralized services, because Storj enables multiple nodes to deliver files simultaneously in parallel from the edge, and no single device or network is a bottleneck.
Economics – Storj can be incredibly economical, since it leverages the vast amount of unused storage capacity that is out there. Most farmers report that they don’t spend anything incremental on equipment, bandwidth, power, or people. This allows us to provide a very economical solution to customers while providing healthy margins to farmers, partners, and Storj itself.
What’s the basic business model?
Right now, the company’s model is really simple and compelling. People who want to store data (our customers) pay a certain amount for storage and bandwidth as they use it. We give the majority of that money to our farmers. We keep the remainder, which is often split with partners who help us build demand, such as user-facing file management companies or many open source projects. In many ways, we are like Airbnb. Airbnb is the largest hotel chain, but doesn’t own a single hotel. We can be the largest storage company, without owning or operating a single data center. Storj is a software solution to what has traditionally been a hardware problem. One of the things I like best about the model is that it provides meaningful monetary rewards for everyone. Most customers will save significantly over traditional alternatives. Farmers will be compensated for sharing their excess storage capacity. Partners can create a new revenue stream for their business. For open source projects, this monetization opportunity can be huge, giving us a way to help them monetize in a way that the traditional cloud companies often don’t. For example, for open source partners like FileZilla (and many others waiting in the wings), Storj has emerged as one of the most effective means to create financial sustainability.
Of course, one of the best parts of Storj Labs is the team. I was incredibly impressed by the founder, Shawn Wilkinson, who starting working on Storj in his dorm room at Morehouse College, and bucked the “Silicon Valley” model by building a great company, on a shoestring, in his hometown of Atlanta, through bootstrapping rather than large infusions of VC cash. The team that he has built around him is incredible: grounded, sophisticated, and dedicated to customers and the broader community. And, that community is already amazing.
What about the cryptocurrency? What about the crypto bubble?
Blockchain is about more than just cryptocurrency. Decentralization is bigger and more important than any single technology, such as blockchain. And, Storj is fundamentally a decentralized storage, computing, and applications company.
We do have our own utility token, STORJ. Farmers are compensated in STORJ based on their contributions of storage capacity and bandwidth to the network. Prices for customers and rental rates for farmers are quoted in dollars. And, customers can pay in their choice of dollars or STORJ. Today, the token serves a medium of exchange within Storj’s closed-loop ecosystem of farmers and developers, but we’ve designed the system to be flexible.
We believe that the best thing we can do for the value of the token is to focus on building a large ecosystem of customers, farmers, and partners. That having been said, we are also committed to being leaders in transparency and governance, and providing as much utility and value as possible to the users of the Storj network.
What are your priorities as executive chairman (and interim CEO)?
My priority is to build a platform and company of lasting value. Over the next year, my top priorities will be:
Setting the foundations of a long term, sustainable business. This will include working with the team both to set a long-term strategy, mission, roadmap, and business/financial model, and to build a strong organization, organizational model, culture, and set of values that can deliver excellence and stability in the midst of hypergrowth.
Delivering excellence to customers. I’ve spent time throughout the IT stack, and I know that the bar for security, reliability, and performance in storage is higher than almost anywhere else. While I am excited about what Storj has achieved to date, and believe that the fundamental technological model is sound, we will need to deliver sustained excellence over many years to earn the trust of customers with the most sensitive workloads.
Creating value for the broader community. If we are to be a valuable platform and build a fundamental layer for the decentralized internet, we need to help build both technical and economic value for the broader set of people who both build and rely on our platform. In addition to delighting customers, this means ensuring that it is easy and financially rewarding to be a farmer, to build applications on Storj, to build user-facing businesses that leverage the Storj network, and to contribute to the Storj projects. I also think we have a unique ability to provide a viable monetization model to open source projects that leverage and generate demand for the Storj network (something that often isn’t possible with traditional, centralized cloud computing models).
Why Executive Chair vs full-time CEO?
Executive Chairman and interim CEO is the right role for me and Storj Labs today, given my availability and Storj’s near term needs. Now is a critical point in Storj’s lifecycle, where I can make a large impact in establishing strategy, building a strong organization and culture, putting in place key partnerships, and establishing a sustainable and healthy business model. Although I am not able to commit to a full-time CEO position at this time, I am committed to a highly-involved, operational role as Executive Chairman and interim CEO of Storj Labs.
This is a hugely exciting journey. I am excited to be sharing it with the Storj team and the broader Storj community. I’m expecting great things in the months and years ahead!
Read about the genesis of Storj and its evolution in “The Next Generation of Storj” by Shawn Wilkinson.
* If an individual node has a failure rate of p, and there are n completely independent nodes in a system, the chances of them all failing is something like n^p. So, if your network has nodes that have a 20 percent failure rate, but 10 have to fail for the system to fail, your system has something like 1- (.20)^10=99.9999% reliability. Farmers have incentives to have much higher reliability than in this example, and far more than 10 nodes need to fail for data loss to occur.