Centralization Strikes Again - May 2021 Edition
“Arguing that you don't care about the right to privacy because you have nothing to hide is no different than saying you don't care about free speech because you have nothing to say” — Edward Snowden
Where have centralized technologies, including those who limit our access to it, failed us this month? Let’s take a look at our May month-in-review:
Timeline
May 4 - Robinhood goes down amidst Dogecoin and Ethereum rally. Source Source 2
May 4 - Nasdaq Inc. exchange computers suspend broadcasting prices of Berkshire Hathaway stocks due to ‘number too high’ (precisely $429,496.7295). In a similar situation IEX Group Inc. stopped accepting orders of Berkshire stocks “due to an internal price limitation within the trading system.” Source
May 6 - Incoming SEC Chairman Gary Gensler suggests bringing tighter regulatory oversights on cryptocurrency exchanges. Source
May 7 - Sci-Hub creator Alexandra Elbakyan shares a delayed notice email sent from Apple, in which Apple was subpoenaed by the FBI two years earlier in request of her account data. Source
May 9 - As Elon Musk made his appearance in SNL, Dogecoin prices plummeted, causing issues with trading on Robinhood. Source
May 11 - LinkedIn freezes account, removes comments of prominent critic of China. Source
May 12 - Researchers find critical security flaws in Australia’s electronic voting software: "Secretive, unverifiable systems like the ones used in the ACT 2020 election, make it relatively easy to change the recorded list of votes cast, in a way that observers cannot notice.” Source
May 13 - Microsoft announces it will be shutting down its Azure blockchain. Source Source 2
May 13 - Binance comes under investigation by the U.S. DOJ and IRS for money laundering and tax offences. Source Source 2
May 15 - WhatsApp’s new privacy policy — which gives them the ability to share user data such as phone numbers, IP address, location data, and device details with their parent company Facebook — officially kicks in. If users continue to refuse to accept the policy, app performance is said to imperceptibly deteriorate until completely unusable. Source
May 16 - Twitter suspends Spanish far-right politician Francisco José Contreras for controversial tweet. Source
May 18 - China bans financial institutions and payment companies from providing services related to cryptocurrency transactions. Source
May 19 - Binance suspends Ethereum and ERC-20 withdrawals while Coinbase, Kraken, and various other centralized exchanges and data providers all go down amidst market-wide panic. Source Source 2
May 24 - Paul Graham reports Vox stealth edits on articles debunking conspiracy theories of COVID-19 lab leaks. Vox has since retroactively added statements of acknowledgements. Source
May 26 - Last day for all social media companies operating in India to comply with a local regulation that mandates all firms, including Facebook, Instagram, and Twitter, to offer traceability on data or face criminal liability for content deemed illegal. Source
If any newsworthy bulletins about centralization have been missed this month, please reach out via email or Twitter - we will get it added! There’s just too much to keep up with, and we’d deeply appreciate your contributions (credits will be attributed!) as we continue to track the failures of centralized models.
Op-Ed
How Filecoin works, and why it's an existential threat to the Big Tech business model
Quick intro; I'm Hunter, and I work on ChainSafe's Forest team. Forest is an implementation of the Filecoin Protocol, written in Rust. Forest is meant to be a drop-in alternative to Lotus (which is written in Go), and will take advantage of Rust's speed and correctness to build a more resource-optimized and stable Filecoin node.
I'd love to write more about some of the really interesting work we've been doing on Forest at some point, and as we approach our 1.0 after the security audit we're currently working through, I intend to help provide more context on Forest itself.
For now, however, I would like to address a more general concept relating to Filecoin: how it's kept secure, and how this relates to overthrowing the corporate cloud computing hegemony. The technology Filecoin is using is arguably one of the most sophisticated and rigorous approaches to solving the problem of decentralized storage. In some ways it's still quite experimental technology, but its importance should not be understated.
As opposed to Proof-of-Work, such as in Bitcoin mining, storage mining makes storage (the capability to store data, such as files) the scarce resource that anchors FIL— the monetary asset— within reality, via something that is provably scarce. Computers generally are not much grounded in reality otherwise; for example, a number could be held in the space of 512 bits that is much larger than the number of atoms in the known universe. It's hard to disprove someone saying they have a million bajillion coins if they do not rely on a source of scarcity to secure the network.
Filecoin uses both Proof-of-Replication (PoRep) and Proof-of-Space-Time (PoST), to facilitate storage mining. For the first, PoRep, the process of proving replication is often done in 32GB or 64GB sectors, which are required to be prepared beforehand via a process known as sealing [more on this can be found here: https://spec.filecoin.io/systems/filecoin_mining/]. This proves that storage capacity available on specific hard drives are committed to the network. This is like preventing "double spend", but in this context, preventing "double store."
PoST is then used to prove that the storage capacity has indeed been available over a certain period of time.
Committed capacity is just a sealed sector that isn't storing real client data, but is capable of doing so. When a client makes a storage deal with a specific miner, a sector is prepared that bundles their files, then sealed once more by upgrading their original committed capacity, and over time, proven that they were made available to the network.
Additionally, money (FIL) is staked with this committed capacity, so that there's financial incentive to reliably maintain availability of stored data to the Filecoin network over time. If a miner fails to prove data availability for too long, their stake is slashed.
This then allows for a few things. Storage miners are able to earn money as soon as their storage is prepared and made available. This allows infrastructure to be built before it can be used, which helps in building out a more decentralized network. This will allow for "Uncle Jim" with a server rack in his basement on the other side of town to run a profitable and decentralized competitor to traditional cloud data centers within a broader shared storage market. With Filecoin, a user's data can be stored across many different miners, and since it's prepaid for a specific amount of time (currently up to a year and a half), this competes directly with cloud-based "storage as a service" providers that will delete or restrict access to your data if you miss a month's payment. And because miners get paid for mining regardless of demand, oftentimes they'll store data for thousands to even billions of times cheaper than cloud providers. [A listing of storage miners and their fees is available here: https://file.app]
And should you ever decide to leave the cloud, not only is there friction from changing interfaces (APIs and UIs) and software for accessing and storing your data, but the very act of liberating your data can be prohibitively expensive if you have enough of it.
Now, there's certainly friction right now for using and accessing Filecoin, due to its relative immaturity. It's in the process of replacing an industry that has had more than a decade's head start. But that's expected to improve over time. And at least it's the same friction no matter where you want to store your data. Filecoin is a network designed to provide the underlying infrastructure that can be used to compete with major players. It gives us the tools to take back the web from the handful of corporations that control it, and make the web look closer to the internet that it's built upon. The difference between the web and the internet is, the web has clients and servers, whereas on the internet—which operates on a layer below— the packets are routed from IP to IP regardless of what they're meant for. And a web where every browser is also a server, such as what IPFS makes possible, is a web that is inherently more efficient, reliable, and trustworthy. A centralized web is designed for scale rather than efficiency or reliability, because it's meant to to be run under the ownership of only a handful of organizations, and that's something they can never be expected to solve on their own.
Admittedly, from a technical perspective, Filecoin storage is also somewhat inefficient. Not only is a replication factor of about 10x suggested, but a write amplification factor of roughly 10x is also needed in order to seal data. In addition, sealing and unsealing data takes many hours. However, by pairing this with IPFS and Filecoin retrieval markets, Content Delivery Networks (CDNs)— which often host duplicate copies of data in different geographic regions— can now be organizationally decentralized. Some CDNs, including Akamai and CloudFlare, have taken down a large percentage of the web whenever they run into trouble. IPFS and Filecoin are distributed systems designed to obviate this entire problem.
So, seeing as the integrity of the web can be compromised in many ways, the only way to solve for that is to make it so ownership is returned to users and content contributors, who are the real stakeholders in the web. Filecoin's compensation model reduces perverse incentives in traditional cloud storage models where a single company is trusted to store and serve data. Filecoin is building the technology and infrastructure required to get us there. It's still early days yet, but we're making good progress, and as Filecoin is made easier to use and more commonplace, there are certainly some very exciting things that lie ahead.
Ultimately, the only reason Filecoin won't disrupt the corporate rent-seekers is just that it's hard to use. That won't be the case for very long, and as such, the days of the corporate stranglehold on cloud computing are numbered.
Thank you to Hunter Trujillo (from ChainSafe’s Filecoin Forest team) for his contribution to this month’s edition of our newsletter.
If you would like to contribute to the Op-Ed in next month's edition of Centralization Strikes Again, then please email tim.ho@chainsafe.io. We would love to hear your thoughts on how we can leverage decentralized technologies as a solution to the many problems centralized systems present for humanity!
Like all great things, our newsletter starts humble and, well... unpolished. If you have any feedback, please don't hesitate to send them my way.
Learn more about ChainSafe at https://chainsafe.io/, through our Medium, via Twitter, or by visiting the ChainSafe GitHub.
❤️ Thank you for reading! ❤️