Its players have so far spent over 15 million USD buying collectable cartoon cats in two weeks since launching on November 28th. The cartoon cats can be bred to create new, genetically unique, and tradeable offspring. The game has proved so popular that the transactional aspects of the game now account for over 10% of the network traffic on the Ethereum blockchain. This matters as with increasing load, transactions become more expensive to execute quickly. Not great if your service relies on timely transactions on the Ethereum blockchain.
This reminds me of the free-riding problem in the early days of P2P file sharing platforms. Not quite the same problem, but the impact is analogous – a small slice of the user base makes disproportionate use of a shared resource rendering it less useable to the rest. Arguably though – that is just the symptom. The root cause being the limited transaction scalability of the underlying platform.
The inflationary effect of rising demand
In a resource limited platform one would expect that disproportionate use is discouraged through transaction fees. CryptoKitties shows us that this strategy can be ineffective when there is speculative value attached to transactions, fuelling seemingly irrational demand and corresponding load on the blockchain. Using transaction fees alone as a deterrent renders the platform unusable to services that rely on time-sensitive low-value transactions.
Another approach adopted by ISPs to manage an analogous problem in broadband service provision is to enforce a “fair usage” policy. Instead of using pricing as a deterrent to overuse of a limited resource, disproportionate consumption is curbed through usage throttling. Most users are satisfied as pricing is unaffected, but peak users feel restricted. For many blockchain use cases such an approach is not attractive as it puts a cap on service growth. Such a policy is also more difficult to enforce on a blockchain. It is also unattractive to those entities running the blockchain fabric as they get rewarded for verifying transactions and higher transaction fees mean more lucrative rewards.
The problem then is that with a limited resource (transaction throughput), transaction speed becomes a valuable commodity, and unless the cap is lifted, the price tag may rule out use cases that are inherently low-value, but still reliant on speed of transaction execution. The holy-grail solution is of course to make the platform inherently more scalable. Blockchain transaction scalability is a recognised problem and there has been a good amount of vision around it in the Ethereum community (see also this). This is not a trivial problem and chances are that services willing to rely on the Ethereum blockchain (or any other public blockchain for that matter) will be exposed to the current limits for quite some time.
What does this teach us?
It may sound harsh, but public blockchains are not yet quite ready for time-sensitive mission critical use cases that cannot justify inflationary transaction fees. But this does not necessarily mean the end of decentralised trust.
Like the P2P sharing platforms that preceded them, Ethereum and Bitcoin blockchains have proved a point. Extracting disproportionate value from serving as an incumbent trusted third party is increasingly difficult. Also some services are inherently more palatable to users if no single entity has (or is perceived to have) a monopoly on the service.
While the technical realisation will change, possibly quite substantially, the value chain disruption is likely to stay. The P2P sharing platforms at the turn of the century curiously have led to the rise to centralised platforms such as Spotify and iTunes. Technically the opposite to the P2P platforms that started the transition, but nevertheless fundamentally disruptive to the incumbents.
Hybrid crypto chain platforms to the rescue
While much attention and publicity is concentrated on the public blockchains, some use cases are better suited to run on platforms that are less exposed from falling hostage to speculation and other undesirable effects.
For many commercial applications, consortium blockchains offer many of the benefits associated with blockchain as a decentralisation platform, while placing barriers on disruptive transactions. Often shunned by blockchain purists, consortium blockchains provide a hybrid between fully distributed public blockchains and traditional centralised trust.
In consortium blockchains, the trust base is distributed across partnering organisations or even willing competitors. That way your time-sensitive, mission critical, and potentially privacy sensitive transactions do not need to contend with CryptoKitty transactions. For applications that require a broader (globally distributed) trust-base, blending between layers by carefully selecting which transactions need and should interact with the public blockchain is key.
We are already used to this “trust hierarchy” in the everyday life. National governments issue passports as a proof of identity and nationality, but most of us have a plethora of other ways of identification that we use in other contexts such as accessing physical assets (office space, vehicles…), financial assets, and digital resources. It is also how much of the internet fabric is run, with private (or semi-private) infrastructure instances used for local context (e.g. local DNS domain) all the way up to shared infrastructure (global top level domain, and root servers).
CryptoKitties is probably the closest that Ethereum has so far got to becoming a victim of its own success. At just under 1 million transactions per day, the Ethereum transaction volume still dwarfs in comparison to VisaNet’s 150 million transactions per day. Once that limit is reached (unlikely in the near future) demand will drive new limits. Meanwhile you can still take advantage of blockchain as a trust platform without being hostage to speculation… and CryptoKittens.