Sia: Traveling for the Destination, a Long-term Investment (Part 2)

Sia wants to be the hard drive of the Internet. Underneath the hood, Sia is a cloud storage project, like Dropbox, but it is decentralized and powered by both a blockchain and its own cryptocurrency.

In this blog series I look at the Sia project from various angles. I provide pointers and share some thoughts. None of this is investment advice, though: research yourself and use your own head.

Parts of this series: (1) Intro — (2) Long-term — (3) Mid-term — (4) Competitors — (5) Investing — or Read more about Sia


So, this journey with decentralized cloud storage at the end is happening. Woohoo! You might love this destination—but there are many buses going there. Should you join the Sia bus or another one? Here's a couple of questions we'll ponder:

  • Is it worth going? Three days full of hazards to reach a 7-11 shopping experience… maybe not your idea of a great trip. Knowing the destination, will it have been worth it?
  • Who is on board? You wouldn't want to have a notorious alcoholic as your designated driver for a road trip. So let's talk about who's embarked to revolutionize cloud storage.
  • How's the engine doing? The car needs to be sound, unless you want to spend your vacation inside a repair shop. How're the game mechanics and the tech doing, and are they supporting the goal?
  • Why is the team going there? Learn about the Sia Ethos.

Related questions we'll discuss later:

  • Woohoo! We've made it…wait, who are those guys? No fun, if someone else is there first. So, who's in competition with Sia? Find out in part 4.
  • How can you join the journey? You're not going to be in the driver seat. Will you be in the back? In the trunk? Look at a live video feed? Basically, what options do you have to join in the Sia journey and profit from its success? Read part 5 on how to join the journey by investing money or time.

If that was too much text for you already, do jump to the Knowns & Unknows.

Is it worth going?

Consider the situation where after many years you finally succeed in becoming the premium provider of second-hand match boxes in Fiji. Why, then, it seems congratulations are in order! But, you're not making much of a dent. So, here's a fair question to ask: Let's say Sia becomes that hard drive of the Internet. How big is that? Intuitively, it's bigger than second-hand match boxes in Fiji. But, how much bigger? Is it also bigger than the market for second-hand, first-hand and any other kind of match box in the world?

$25 billion seems to be the size of the cloud storage market today. (“seems” and “may” because I don't trust reports all that much whose data sources and methodology I can't inspect.) That's big. The same reports suggests, it will only get bigger, reaching 92 billion in five years. That's because infrastructure gets moved to the cloud.

Also, humans are generating more and more data. More people have access to the Internet, and the amount of data per person per time grows simply because we have better connectivity and more cameras. The IoT is growing, with all its sensoring. AI is producing learnt data from the gathered one. More unstructured data everywhere. In 2012 we allegedly generated 2.5 quintillion bytes per day.

As for growth, it seems our storage needs are growing exponentially, not linearly, even though you can't believe every fancy marketing statement, surprise. At the same time, storage size for hard drives have stagnated in the last years—we have our data in the cloud. Still, storage capacity has not grown exponentially. Not seeing Petabyte drives anywhere.

I've no idea how big the world-wide market for match boxes is. But (cloud) storage is big and it will only grow for the foreseeable future. If Sia can carve out a chunk of that projected 92 billion, it's got a shot at becoming a unicorn, or at least a many-million-dollar pony.

Who is going?

Alright, the destination is great. But who is driving? Who is reading the map? Has anyone done such a trip before?

First, the company developing Sia is called Nebulous, and partly nomen est omen applies: The domain nebulouslabs.com redirects to Sia, and the closest thing to a team page I found is the Nebulous page on the Sia wiki, which at least seems to be updated regularly.

Nevertheless, good news for paranoid You in a world full of ICO scams: Nebulous is very likely a real company! (Not necessarily in the legal sense, I haven't checked.) The Sia project has been around since 2013, its software and network maintained and improved for as long. A Boston Globe article presents Sia. (See Part 3) for the controversy around that.) And, David Vorick, one of the founders, has been discussing Sia for years.

David and Luke seem to be the founders. Here some pointers to their online life.

David Vorick, co-founder, lead developer. Twitter, GitHub, on the Sia blog, on LinkedIn, on Reddit.

Luke Champine, co-founder of Nebulous and Sia core dev. Personal homepage, on GitHub, on LinkedIn

Other team members are: Ava Howell as Sia UI and Sia Core developer, Christopher Schinnerl as Sia Core developer, Steve Funk as Community Manager, and Zach Herbert as Vice President of Operations.

One obvious problem is the missing skill in terms of sales and marketing. In fact, until recently when Steve Funk and Zach Herbert joined, the Sia team consisted of developers. That's already a cliché of a reason why tech startups fail. A cliché, because it's so true. And yet it's of course not a necessity that Sia fails, as long as they recognize it and take strong counter-measures.

To make another point, here are some rough numbers to ponder, for the individual team members in the order above, with data from the team page, LinkedIn or other pages linked above:

  • Years since high-school/college: 3, 3, ~5, 0?, 10, 0
  • #Jobs before: 1, 1, many, 0?, 3, 0?
  • #Startups before: 0, 0, 1-?, 0?, 0, 0

This isn't perfect science, so assume everything is off by 1 and some numbers may be off by a lot, but generally, the team is very young and especially the leadership team hasn't done a crypto startup or any startup for that matter before. (Luke's 4 months at Akamai are useful, but not a deep dive.)

That's worrisome, and even magnified by the advisors, whom the Sia team relies on. I didn't find any. I'm aware that flashy advisory boards are regularly used just to promote ICOs, and there seem to be bogus advisory boards, too. The original idea of advisory is sound though: Outside and strategic advice from someone experienced in what you do.

My concern is two-fold: The Sia team is young and inexperienced, and they need good advice. (Like I do, maybe I'm just projecting?) And maybe they know this and have this support structure set up. If yes, then they don't communicate that particularly well. There is another plus of a known advisory board: Experienced “public figures” with a reputation to loose attach their name to the project, and that lends it legitimacy and it provides these advisors with an incentive to help the project succeed. They, too, are then … on board. (Amazing pun, I know.) I think the Sia project is a bold one and it only stands a chance with a lot of outside help.

Now that we know who is going, let's talk about incentives. For a project this size to succeed, we need to understand who profits how from Sia's success. For Nebulous, the answer is a 3.9% fee from each successful file contract. As explained in this reddit thread, it's about 10% of what a client is paying. The mechanism to collect that money, Siafunds, is the second cryptocurrency in Sia, and you can buy it, too, and that's why I'm explaining it in more detail in Part 3.

Engine / Technology

For some road trips, the journey itself is the goal. In contrast, the Sia journey is about reaching a destination, and if you want to reach one, you need to spend your time on the road, not in a repair shop. And so, we need to have a look under the hood.

In this part, I'll focus on architectural choices, which are hard to change. There are rough edges for early adopters at the moment, too, but I'll discuss those in part 3, as I expect the Sia team to address them eventually.

Blockchain, why?

For any blockchain-based idea, the most urgent question is: Does this need a blockchain? (It's urgent because it really kills most projects at inception. Look at your current most favorite ICO, ask this question and don't come crying to me.)

So, does Sia need a blockchain? And why?

To remind ourselves, a blockchain keeps a history of records secure and public. Because it's public, you and anyone can verify that it's not been tampered with and therefore you can trust it's records. Sia needs to establish this kind of trust because it wants to be decentralized: As client, you rent storage space from someone you've never met, and never will, on a server you've never seen. If you're storing something important, say pictures of your first donut, then you want to be reasonably sure that your data is safe and available.

Wait, why does Sia need to be decentralized? Well, decentralization is Sia's main selling point over existing, centralized providers, such as AWS S3. I'll discuss the merits of decentralization below. So, let's assume for a moment that decentralization is a good thing. Who needs to trust whom about what, and how does the blockchain help?

Let's say Tommy wants to rent 5 gigabytes for one month from Mia for 920 money. There's a lot of trust necessary here: Tommy needs to trust that Mia will keep his 5 GB worth of files around for an entire month. He needs to trust that he'll be able to access them any time he wants. And Mia on the other hand needs to trust, if she does all that, that she'll get paid.

So, here's how the blockchain helps to establish that trust: A storage contract documents that Mia is providing 5 GB storage to Tommy for one month. It's public, there is no arguing. Furthermore, over the duration of the contract, Mia has to regularly prove that she still has Tommy's files around. Those storage proofs also become part of the blockchain.

Mia can trust to be paid: As part of the contract, the blockchain reserves 920 of Tommy's money. If he does not have that, there is no contract. The network can ensure payment, because it brings it's own “coin”, Siacoin (SC). When the month is up, the network pays out 920 SC to Mia's wallet.

The blockchain also serves as Sia's yellow pages: Hosts can advertise their storage capabilities and offerings in the blockchain, as part of some “free-form space” in the block. This means there is no central service necessary to find suitable hosts, either.

Technically, the Sia blockchain is based on the Bitcoin source code. (The Sia blockchain itself is not a fork of the Bitcoin blockchain. Sia merely copied the software code in 2013.) One implication of this choice is that Sia relies on proof-of-work to secure the blockchain. After copying the source code, it was extended to include new transaction types for the above contracts and proofs.

Running your own blockchain technology does not come free. (That's why Storj did not.) You need to maintain it. Read in the FAQ why Sia wants its own, instead of using, for example, Ethereum.

Storage contracts

Let's look at storage contracts in more detail, as they are the basis for renting in Sia. You decide up front how much space to rent for how long. Reputation for hosts and payouts are coupled to completed contracts. When thinking about storage, we often think about files. Sia decided to go with storage space instead, and it's important to understand the implications.

First, it's good to note that Sia does know the concept of files for storing, even though its in the Renter API. That means you can rent a Terrabyte, start uploading files, and the Sia client will take care of distributing them over the hosts you've a contract with. Sia uses erasure codes for that. Read Vitalik Buterin's blog post for technical details on erasure codes.

So, now you know how much space you want—how do you know how much to pay for it? You don't. Instead, you specify an allowance in Siacoin that should be consumed per contract period (currently 3 months). Once you lock that money in, your client negotiates a bunch of storage contracts with hosts, after which you're ready to go. Or in other words, rather than telling the system how much storage you need and accepting the price for it, you give it a budget to work with. Does this sound complicated? To me, it does.

Part of the problem is a user experience bug, which can be fixed: Instead of starting with a budget, you could specify the number of GB you want. The client would then look at available hosts and their pricing and give an estimate on how much budget would be needed, possible with a safety margin. You could then accept this budget, and the client would proceed as it does now, negotiating the contracts.

The other part can't be fixed easily though: There is no price guarantee. Since hosts can become unavailable, rented out, or simply not accept contracts, the network cannot guarantee any price to the user until the final tally will be drawn up. That's because the Sia network does not guarantee a price, see the Upload files section in Sia's Getting Started blog post.

Since you rent storage space by setting a money allowance, getting more storage is easy: Just increase the allowance. Extending a contract is easy, too: It actually happens automatically. Half-way through a contract (currently after 1.5 month), the client will try to auto-renew it.

As a side, Sia sometimes calls its file contract “smart contracts”. When reading about smart contracts nowadays, I think a common association are Ethereum's smart contracts. So I thought it's good to compare them quickly, as Sia contracts are not the same as Ethereum ones. Both are open-source: you can read and understand them. Both guarantee something, and in Sia's case this is the trust mentioned above. The first major difference is where the contract code is stored. For Sia, that is simply GitHub. The configuration for this code (who rents how much for how long from whom) is in the blockchain. With Ethereums smart contracts, the code is in the block-chain, too: Both parties can create custom ones. That is strongly related to the second major difference, namely what the contract can guarantee. Sia's contracts are specific to renting storage space from hosts (DSC for domain specific contract, anyone?). Ethereum's smart contracts on the other hand use Solidity, a programming language, which unfortunately is turing complete. Since anyone can create a new smart contract by putting code into the blockchain, you can do whatever the runtime environment of Solidity allows you to. That's way more powerful. Nevertheless, both are smart contracts in the sense that they provide guarantees, and don't require you to trust the goodwill of the other party.

To me, Sia's contracts feel as heavy weight as renting an apartment. With an apartment you know how many rooms you'll need, you can commit for a long time and those things are unlikely to change any minute. That's not necessarily true for storage, where a fine-grained pay-for-usage with a post-paid model is common (see S3 pricing). Overall, I'm not convinced that individual contracts for space between client and host, driven by a budget are a great abstraction. I guess that remains to be seen. :)

Siacoin

Siacoin, Sia's cryptocurrency, is currently mostly used as an investment or day-trading coin. (See part 3 for the current state of the network.) If it is to serve its purpose to secure storage contracts, it needs to become stable: Hosts and clients alike want to know what amount of money they commit to, at the moment for months. That means, in the end there should be little volatility and slow, or at least clear trends. So, how is Siacoin constructed?

As mentioned, the blockchain is based on Bitcoin. In contrast to Bitcoin, however, there is no limit to the number of Siacoin. There is an infinite supply; Siacoin will be mined forever. Difficulty will be adjusted as with Bitcoin, but over very long periods, there would be insane inflation.

A mechanism that will counteract this is “proof of burn“, which at the time of this writing is still on the medium-term roadmap. What is proof of burn, and why do we need it?

Imagine a hacker kid: smart, weirdly enough without acne, socially pleasant, hates diet coke, and just needs to hack. That kid creates a whole bunch of Sia hosts and adds them to the network, and at the beginning those hosts play nice. They then accumulate good reputation and get more and more file contracts. At the end, some of your files are only stored on hosts of the hacker kid. Now, the hacker kid has you by your private parts, and giggles in her basement. How can you prevent this situation?

The hacker kid can only add that many hosts because it's cheap. If the kid would have to spend actual money on each host, even a bit, then a couple of hosts would be possible—but not many thousands. And that's where proof-of-burn comes in: To claim a file contract, a host will have to prove that it burnt some Siacoin. That means spending money, and that means the hacker kid is out, unless it's an insanely rich hacker kid and then we've got ourselves a different kind of attack(er).

For the purpose of our Siacoin discussion it's enough to realize two things: First, all the time Siacoins will be taken out of circulation. Second, depending on the exact numbers, this could result in deflation, too.

By adjusting mining difficulty and proof-of-burn, Sia has two tools to control the number of coins in circulation.

Another important detail with cryptocurrencies is the ratio between circulating and outstanding coins. Many cryptocurrencies suffer from heavy premining, which holds the risk of hyperinflation. According to Sia's announcement, about 30 million SC have been premined by the devs. I could not confirm this, but in either case it does not matter too much, as there's currently a circulating supply of 31 billion coins. Nebulous is not going to permanently crash the price with those 30 million coins. :)

Challenge: Scaling the blockchain

Since it's essential for finding hosts, contract formation and payouts, a major challenge will be to scale the blockchain.

Sia is based on the same technology as Bitcoin, so it's fair to ask how it will address the scaling issues that are haunting Bitcoin: throughput (number of transactions per minute), high mining fees and confirmation time. Storj, originally based on Counterparty, had to switch in part because of these scaling issues and lost momentum because of it. In contrast, Sia is currently doubling down on its blockchain, by securing it with Obelisk ASIC hardware miners, see the part about Siacoin above or part 5. These problems lead the author of this blog post to believe that the market for block-chain based cloud storage is actually quite slim.

It's good to note that Sia is well aware of the confirmation problem. With a block being mined every 30 minutes, by error, and no extra magic, it takes an hour to secure a file contract. Cutting that down to 5 minutes is already on the roadmap.

These scaling issues of course are not only a problem of merely Sia or even Bitcoin, but of virtually any blockchain whose tokens are used as a currency of sort. This means that one possible strategy is to just let the rest of the community figure it out, and then copy the solution. :)

Still, the challenge to maintain Sia's own blockchain over time is a formidable one.

Challenge: Self-Maintainence

Any useful decentralized system has a single point of failure, from the client's point of view: the client itself. I think it's the network's responsibility to take care of technicalities; that's what it's being paid for. The client (software) really should be quite stupid and responsible only for the most fundamental decisions (what to store? how secure?) and information (client ID, encryption keys).

For Sia there are specifically two issues I see quite critically:

  • Availability: If hosts get lost, the client currently must be online (and I think even have the file locally) to spread it to new hosts and retain it's chosen replication level. I think it should be the network's responsibility to retain a chosen (and paid for) replication level.
  • Auto-renewal: If a contract is half-way through, the contract is open for auto-renewal. But only as long as the client is online. Again, I think this should be the network's responsibility.

In this area there are some design flaws, which I'm not sure about:

  • File contracts: Instead of making a contract with the network, the client makes it with individual hosts. That means it needs to find individual hosts, vet them, and then make contracts. I think this is too manual and gives the client too much opportunity to screw up.
  • Currently, the client remembers where which file part is stored. That means, it's not enough to have the encryption keys and wallet backed up; you need to have that Sia data directiory, too. If the hard disk of the client goes banana and there is no backup of that data directory, all files are inaccessible forever.

Challenge: Reputation system for Host scoring

What client's ultimately trust is the Sia network's ability to keep its promises, starting with the basic promise that a file will be accessible for the time that was paid for. Malicious, overgreedy or incompetent hosts are a problem for this promise. To distinguish between those hosts and the quiet-running, well-maintained ones, Sia is using a reputation system.

Hosts are ranked against each other using a score: the higher the score, the higher the host is in the ranking and the better the chances of getting file contracts. Factors that influence a host's score include uptime, pricing, age and version of the Sia software running. For details see the excellent Wiki article on the topic.

Conceptually, that's a straightforward solution: Good hosts earn themselves a better reputation than malicious or incompetent ones, so they get more and better contracts, making it less profitable to not play nice.

Reputation then is money. And that means lots of hosts, even the good-meaning ones, will try to game the system. Keeping the reputation meaningful is a constant arms race, and a big challenge for sure.

Challenge: File Access QoS

File contracts and proof-of-storage incentivize the host to play nice and keep the files around. Since the storage proofs are part of the blockchain, failure to do so is very visible. That's an awesome feature, so much so that people don't mind loosing money over it. Here's the catch: There is no such mechanism for downloading and uploading files.

Essentially, the problem is that it's possible for the host and client to lie about whether they sent or received a file. At the same time downloading files cost the host bandwidth and the client Siacoin, so both parties are incentivized to lie. It's easy to lie for either, because from a technical point of view the network has no visibility into the file transfer.

From a game-theory standpoint, the problem can be fixed using a reputation system as mentioned above. This assumes that the majority of hosts and clients are honest. It does not prevent, however, that a particular host or group of hosts discriminate against a particular client, and vice versa. It's not obvious to me how to fix that. Possibly, by lodging complaint transactions against the host to the blockchain, which another host with the same data must confirm.

Challenge: HTTP Bridges

Sia wants to be a general purpose storage system. For example, you might want to send a friend a URL to a file, or stream videos directly from the Sia network. Currently you cannot do this, because there is no HTTP/web interface: a client must speak the Sia protocol.

There are three technical challenges: naming (generate a URL, which then identifies a file), routing (find the best hosts to serve parts of the file) and web-access (requests to that URL end up giving you the file). These problems are solved in a centralized way already: S3, Youtube, they all work. Nebulous could just put up a bunch of servers, and it would work. (That's how Storj does it currently.)

That's not decentralized though; access to the files then depends on a single party, Nebulous. It's not obvious to me, how this problem will be overcome. A competitor, Storj, has opted to abandon decentralization for the time being, in order to make it work. It will be interesting to see how the tension between making it work and keeping decentralization a core part of the system will play out for Sia. (See also the discussion about the Sia Ethos below.)

Challenge: Access control

In an enterprise setting, access rights play a big role. Billing should be managed centrally, while every department may want to manage access to its files. Employees have different levels of access to the department's data bucket, they join the company and leave it again. Sia currently secures files and wallet with a single key. Clearly, this is insufficient for any of the above situations.

A straightforward way to implement this, are bridges: Systems that hold the keys to access the files, and which then have the ability to apply fine-grained controls. Since bridges can be chosen, an enterprise could run its own bridge technology, which uses its internal user access right system, like LDAP.

This is conceptually the same problem as mentioned above: how to build decentralized bridges for HTTP and access control is not obvious.

Use-cases & target market

As part 3 shows, there are still a lot of details to smooth out, but those are details. For the long term, it's more interesting to think about what use-cases Sia will and will not be able to support naturally.

Currently, Sia is working with a prepay renting model. Also, the client forms contracts with hosts, not the network. That means use-cases with a fair bit of long-term planning work well:

  • Backups: From a use-case point of view, a AWS Glacier alternative.
  • Large data storage: Storing large data cheaply, for example gene data, that needs to be analyzed. The data size is known beforehand. Low latency is (maybe?) not critical.
  • S3 alternative for defined target markets: Companies with a local business, where local hosts can be chosen.

Sia may not work so well for situations where demand can't be planned well in advance, either in volume or in location. Also, prepaid may be a problem:

  • Enterprise-customers: They love postpaid, and then pay with a delay. We're running into this at engageSPARK, which is a prepaid platform. Considerably cheaper prices may be the only way to incentivize enterprises and their billing departments to adapt.
  • Fast growing companies: Planning, monitoring and administering contracts to suit the data needs of the next weeks may be far too much hassle for people with already little time on their hands.
  • Apps with trending content: If you've uploaded a cat video to servers in Poland, and suddenly it's the rave in Paraguay, then it'd be really good if the network were to adapt and make those files available on hosts close to Paraguay. Of course, for a long time a CDN can take care of those performance issues. But then, Sia wants to become a CDN. :)

Regarding the market in general, I do think all those crypto-based storage projects overestimate the impact of using the “unused” storage. It's a good idea for bootstrapping a storage network, but as I discuss in part 4 about the competition, I do think that economies of scale favor huge centralized providers.

Regulations & Anonymity

I suspect regulations will be a long-term challenge for Sia, in particular in two directions: Geographic requirements and KYC rules. I'm no law expert, so I'll just share a couple of thoughts, leaving the judgment to you.

Geographic requirements come into play when governments demand that data about their citizens be stored inside the country's borders. To satisfy such requirements, Sia would have to modify host selection to include geography. Unfortunately, with Sia's entirely unregulated and open network, there is simply no way to truly know, where a host is located. Some will see this host anonymity as an advantage, but in this case it will render Sia useless as storage provider.

Know your customer regulations become stricter all around the world, with significant money flowing into crypto currencies. Even with Siacoin being a utility coin, I think it's not implausible that the status quo will be challenged when Sia becomes a major player in cloud storage. Think about all those lost tax dollars—and then tell me, that governments won't care. Again, Sia's host anonymity makes it difficult to know who is hosting what and earning how much. Maybe it's enough for governments to go after the exchanges, which hosters need to extract their profits. And maybe not.

I think the host anonymity, that currently is deeply embedded into Sia, will get challenged eventually. For the success of the project, host anonymity is actually not needed, even if big-brother-wary people like me appreciate it. The problem with identifying and regulating hosts is that this usually requires a trusted authority, which goes against the decentralized, trust-free goal of Sia.

And then, there is video streaming. If Sia will act as a CDN for video, that would mean that the Sia network hosts publicly readable files. They are effectively not encrypted anymore: Anyone has access to them. This may just mean that hosts can no longer rely on the argument that they don't know what they are hosting—and therefore would be liable in the same way that other hosters are. Depending on the jurisdiction, this may mean that such public files would have to be pre-screened by hosters, which of course is not feasible for a private hoster.

All in all, these issues hold enough potential to worry me for the long-term prospect of Sia.

Sia Ethos vs Ship It

One important characteristic of Sia is its rather ideological approach. The ideology of Sia is summarized nicely in the succinctly named blog article The Sia Ethos.

The gist of the Ethos is pure, uncompromising decentralization:

As developers, we have been adamant: no centralization is ever permitted in any of the core software.

Startups need two qualities above all: resilience and speed. In that, ideology is a double-edged sword: With luck, it gives a startup team the faith it needs to power through the hardships of its arduos journey. Ideology can be pure resilience, the dream of a different world.

But it can also rob a startup of that other quality it needs most: speed. The Sia team acknowledges this in the same article:

This makes life harder. […] It means that hosts cannot automatically adjust their prices to match the value of the US dollar […] It means that features our competitors can add with just a few lines of code often cannot be added to Sia at all.

And yes, it does make life visibly harder: What functionality Storj easily adds with its centralized bridge and the promise of future federation (see part 4) is way harder to do for Sia in a completely decentralized way.

An interesting test for Sia is the current A3 controversy. A new hardware miner was built by Bitmain, an entity that can be described at the very least as not very cooperative. In their response to this event, Sia discussed values that are important to them, such as “Doing it right matters”, reaffirming a rather uncompromising stance. See the mining section in Part 5 for details about A3.

For now, the Sia Ethos is a distinguishing characteristic to have in mind. In retrospect, we'll look back and know, whether it was part of Sia's recipe for success, or if it contributed to Sia's downfall. (I'd like to believe the former.)

Knowns & Unknowns

Here my personal take-aways. Feel free to disagree.

Knowns, good and bad:

  • Business idea: The market is huge and well worth an attempt to grab market share.
  • Business idea: Decentralization is an interesting disruption of the market. A blockchain plus coin are possibly adequate tools to create a trust-free market.
  • Business idea: Sia is attacking the right problem size, I think: Decentralized storage is big, without being unwieldy (such as the decentralized computer that MaidSafe is building, see part 4).
  • Team: The team has grit, given that they're still around. For crypto, three years is a lifetime. The team is also terribly young and inexperienced.
  • Team: The team even just in manpower lacks sales and marketing skill. (English speakers: Is there a generic word for manpower?)
  • Competition: Sia has a very well funded competitor in FileCoin, and about a year more to make head way. With Storj, they have a less ideological competitor, whose approach may allow it to be faster to market than Sia.
  • Implementation: Sia's blockchain technology, based on Bitcoin, does not naturally lend itself to fast or micro transactions. Building a pay-for-use CDN or S3-like storage system is hard.
  • Implementation: Sia's pricing is obscure, or at least complicated. Not being able to predict what exactly will happen with one's money is a big down.

Unknowns:

  • Business idea: Customers don't pay for cloud storage. Companies do. The by far biggest unknown for me is if the team will be willing, focused and able to hire the right people for enterprise sales.
  • Ideology: Sia is ostensibly unwilling to compromise on it's promise and goal of decentralization.
  • Sales strategy: I think “we're a platform, we do everything” has rarely worked—Sia needs to solve particular problems for customers, and I don't see a clear strategy to market the platform. Will it be a a backup solution to start with? A CDN?
  • Legal environment: Depending on how the legal landscape evolves, hosters or even Nebulous itself may face significant challenges. While the network itself might survive any attempt to take down individual hosts, or even Sia, the ultimate goal of wide adoption would be severely hampered.
  • Technology: Can they overcome the scaling issues that plague Bitcoin, and become an S3 competitor?
  • Technology: Will Sia be able to provide decentralized HTTP bridges, in other words decentralized public access to the network?

My personal long-term view

Sia might justhit a brick wall in terms of enterprise sales and therefore network size. It will also struggle to overhaul it's blockchain technology for scaling, loosing time along the way. Long-term I'm not convinced decentralized cloud storage in general is a good investment. Since it looks better in the mid-term for Sia, I'll re-evaluate this next summer. I've been known to be wrong before. :)