How close can a centralised company get to having blockchain properties (permissionless, immutable, censorship resistant, etc) without actually using blockchain?
We’ve seen both Ripple and TRON announce 100MM USD funds to invest into the blockchain gaming space. The interest and capital is here. How will the centralised incumbents respond? If both developers and players want to have a blockchain enabled world, they could go with a plethora of protocols that are being developed. But what if the incumbents decide to incorporate blockchain-like properties while maintaining their centrality. What does blockchain enable that the centralised incumbents cannot do within its technical and legal capacity? And can these incumbents get the same benefits without having to use blockchain? In this piece, I will compare properties of Steam’s centralised peer-to-peer game item marketplace with blockchain-based marketplaces for game assets such as RareBits and OpenSea, and discuss what centralised companies can do to become blockchain like without actually using blockchain.
Note: a lot of the discussion here will be explorative. just because a company can decentralise a particular component does not mean that it makes the most compelling business use case or the most financial sense.
Motivations
I’ve wondered how would centralised entities respond to perceived blockchain risk for their business. If decentralised infrastructure provides xyz properties, could the centralised incumbent respond with xyz’ properties, ie something similar? Do consumers want to have Snapchat or do they just want ephemeral video so Instagram stories is an alternative — not exactly the same thing but close enough. Is this something that centralised companies can also do in regards to blockchain?
Centralised Company
For the purpose of this article, I will focus on centralised cloud based companies that are closed sourced and profit-seeking. And more specifically, my analysis only applies to companies that:
- Don’t create software that anyone can deploy and run for free (ie they are not FOSS companies)
- Services require network connectivity (as in cloud based)
- Monetise through advertisements, selling subscriptions (SaaS), or fees
Valve created Steam which is a centralised cloud based marketplace for ingame items. One of the most popular game items or digital assets that are exchanged are skins for the game CSGO. Currently the Steam marketplace monetizes by taking a slice of every marketplace transaction which includes selling of new skins[1] and trading of skins between players. Steam has a monopoly on the marketplace, there are no other ways for players to trade skins without going through Steam.
A blockchain-enabled platform first requires the game items to live on a blockchain, think Ethereum’s Cryptokitties. The platforms built on top such as OpenSea or Rarebits provide both a UI and mechanisms for auctions and trading. Although these two marketplaces are hosted[2] it is possible to have self hosted marketplaces like OpenBazaar.
Blockchain Properties
Blockchain is a censorship-resistant, immutable and permissionless ledger. These three properties reflect the high level attributes of this system. They are the manifestations of a distributed consensus mechanism, applied cryptography and network incentives. These properties as we will see are not exclusive to blockchains.
Censorship resistant — It is expensive for entities to censor an individual on the network.
Immutable — State is append only. No previous state can be overwritten[3]
Permissionless — Anyone can run the network.
- This means the code is open sourced
- And that anyone can fork the network
Micro Properties
- Verifiable execution / State transition rules — There are cryptographically enforced rules on how a blockchain can change from state A to state B
- Incentives to maintain the network — There are incentives for those who run the network on their computers
- Fault tolerant — The system is difficult to take down
- Data sovereignty — Users have control of their own data
Can Centralised Entities Have These Properties?
What Can Centralised Companies Do
Here we delve into potential steps that centralised companies can take to have more blockchain properties.
Censorship Resistant / Anonymity
It is possible to architect systems that have varying levels of privacy from pseudo anonymity to zero knowledge proofs. An exchange implement a zero knowledge system where items may exchange hands and to outside observers they have no idea what trade hands and for how much. Even though trades can be anonymous, a game will still be able to identify what items a user has as those assets must be rendered for the player (though there probably is a clever scheme where this is not the case). Note anonymity is a touchy subject if there is money flowing around in a system.
Immutable
Blockchains are immutable, meaning that once state has been committed to the blockchain, it can no longer be mutated. Companies can hash a snapshot of their data every night with the hash of the previous night and publish the output in a public place such as OpenTimestamps or a newspaper. Then create tools for any user to verify the integrity of the data.
Bellcore Labs hashed their researchers’ work and posted the hash in the NYT everyday back in 1992. Steam could create a data structure for all their customers and their item balances and publish the hash in a blockchain and / or a multitude of public sources.
Verifiable Data
Why trust a platform that a process is truly random or that only one hundred of these special items exist when it can be done with cryptography. Gone will be the days of uncertainty.
Verifiable Execution / Smart Contracts
Although traditional business logic does not adhere to the cryptographic guarantees of smart contracts in the blockchain sense. Is it possible to implement the equivalent for a centralised company. Similar to verifiable data, systems can be built to verify that state transitions have occured and that the output is deterministic and correct. A user can verify that a trade has finalised and within his account there is a new item. And if companies want to get fancy they can use tools such as Ricardian Contracts, which unifies traditional legal contracts to cryptographic contracts.
Incentives to Run the Network / Community Governance
Blockchains allow network maintainers (miners) and token holders to make governance decisions about the future of the protocol. Some protocols let these stakeholders change parameters such as the inflation rate or accepting new protocol updates. Centralised companies have governance in the form of shareholders. They are not necessarily end users and may have competing interests. To avoid this, a platform could try a co-operative model where all players are also owners.
Fault Tolerance
Although 51% attacks on blockchains do occur, for the most secure chains these attacks are extremely expensive and seldom in practice. Even if an attack takes place, as long as a user does not transact during that time period they do not have to worry about a transaction being replayed. Centralised systems can be engineered in ways to be extremely fault tolerant, making it extremely difficult to take them down by either hacking or normal server failure.
Data Sovereignty
Any data a user commits to a blockchain, she can extract. Even on privacy blockchains, as long as the he keeps his view keys, historical transactions can be privately viewed. Centralised companies can do the same, building tools to allow users to extract their own data. Although this is not done normally out of charity, regulations such as GDPR article 20: Right to Data Portability forces companies to allow users export their data.
What Can Centralised Companies Not Do?
Technology can greatly push what companies can do to mimic blockchain properties. However this can only take us so far when it comes to legal arbitrage and fork-ability. Technology might be able to de risk some legal concerns, but it is yet to see if it can diffuse all legal concerns.
Legal Arbitrage
Although companies can perform varying forms of legal arbitrage by setting up subsidiaries and incorporating in various jurisdictions, there are legal limits in what a company can partake in. For example there is no such thing as a legal murder market. Even if the company ran privacy preserving infrastructure and had no clue what was running inside of their servers (perhaps using trusted execution environments), a cease and desist would stop their operation. Because there is a single target to go after, centralised companies have extreme difficulty in evading authorities. At best companies can claim plausible deniability on the use of their technology. But blockchains don’t have to. It is difficult to go after the developers if they don’t run the network. Steam had to clamp down on their API since they discovered third party platforms facilitating underage gambling. Whereas decentralised marketplaces have no ability to censor their users who may be anonymous.
Permissionless Running
Nearly all cloud companies have centralised infrastructure running on premise or in the cloud. Try to name a centralised company that lets users run core infrastructure. Yes apps will often run AI models locally, but that is not core infrastructure. These companies must own their infrastructure in order to keep control of all the data that flows to them. Most marketplaces have a review system, these reviews are curated for the buyer. The precise ordering is not merely sorted by time but instead by quality, relevance and deletion of fraudulent reviews. This requires more than just the reviews for product A, but rather A through Z, detecting patterns across all products. Perhaps innovations in federated machine learning will change this in the future, but until then it will not be possible. This is a problem that plagues peer-to-peer marketplaces. For example OpenBazaar does not have the best curation layer in terms of search and reviewers with personalisation nonexistent. How would they effectively defend against “review bombing” which happens to centralised platforms like Steam? The short answer is that there is no quick solution.
Fork-ability
Because centralised companies cannot let outsiders run core infrastructure, these companies cannot be forked. No one overnight can decide that they dislike Valve’s Terms of Service and migrate not only their history but everyone else's history to a forked version of the platform. Even if the code of the marketplace was open sourced, the data will still be missing which arguably is the most valuable component. In contrast, in a decentralised marketplace anyone can fork the entire network whenever they want to.
Though realistically this may be an advantage to centralised companies, because they are unfork-able there are no distributions to the user experience and political conflicts between different factions.
Perpetual Interoperability (via APIs)
What excited me about decentralised protocols is their permissionless nature, anyone can built on top of the network and no be censored down the line. The greatest fear of building on top of another platform is the cutoff of API access. Horror stories are abound, Zynga bootstrapped off Facebook’s user base, when their exclusive relationship ended, Zynga fell face first. Twitter has done the same with 3rd party apps, cutting off event streaming. LinkedIn too. Why should developers trust that the platform that they build on is benevolent, why should they trust that the platform will not arbitrarily change their APIs or even worse cut off access? Maybe legal contracts could be a stop gap measure, but drive to hoard data is too strong, the more data a company acquires, the better the algorithms, the more sticky the recommendations, why share?
Incentives
Although there are many things a centralised company could do to have more blockchain properties, what are the incentives for them? If you take a company like Valve, what is the forcing factor to be more open (besides regulatory intervention)?
- Providing provable scarcity increases the trust in the platform
- By letting other developers build on top of their infrastructure it increases their distribution effectively for free.
- Having the perception that their platform is more open which increases trust
- Having blockchain properties with a strong brand is good marketing
Advantages to Centralisation
Perhaps centralised incumbents adopting blockchain properties will put them ahead of their blockchain only competitors. Unlike blockchain only companies, these platforms will have a better user experience in terms of personalisation and content curation. Instead of having to see all the items that are available to buy or trade, the platform understands who the player is and what they want to see. Rather than a the uncurated firehose of data that blockchain based marketplaces provide.
Advantages to Centralisation
I suspect that the future will be a happy medium between decentralized protocols and centralised interfaces. Even though Steam has a 100% monopoly of the skin market, what if all the digital assets were issued on an open protocol and rules around transfership and fees were encoded in the asset? Steam as the original asset issuer could program a slice of every single transaction, increasing their revenues. They could operate a centralised user centric data curation layer that provided an improved UX over other platforms and charge for it. And there would be other marketplaces. Cryptokitties has their own auction interface, but kitten owners can trade through protocols like 0x or DApps like OpenSea and Rarebits. Not only increasing consumer choice but increasing distribution.
Centralised Incumbents vs Blockchain Upstarts
- Valve vs Forte
- Ticketmaster vs Tari
- Amazon vs OpenBazaar (distributed tech not blockchain persay)
- Twitter vs Mastodon (distributed tech not blockchain persay)
Only time will tell if developers want to built on top of blockchain protocols or use centralised companies that have blockchain like properties. But the fact that centralised companies can get very close to blockchains should be worrying for their blockchain competitors. I expect that we will end up seeing a future where both worlds end up merging together, with decentralised protocol and centralised entities operating on top of them, who are perhaps even the previous incumbents.
Footnotes
[1] Selling digital skins have zero COGS, effectively zero distribution costs and effectively zero transaction fees. Though a concern is inflating the system and suppressing the price on the open market.
[2] A hosted service in this context means that a user can access the network by simply trying a URL. This contrasts greatly with self hosted services where the user must download software that interacts with the network, requires users to be online, etc.
[3] Blockchains in practice are not as immutable as they say they are. There are attacks that can cause a reorganisation of the chain, however they tend to be very expensive to conduct in Proof of Work systems.