Blockchain is one of those technology shifts that invokes religious zeal, social media fire storms and an unreasoned defence and offence. It is right up there with 5G and covid19 conspiracy theories. The original premise for blockchain centred around distributed data storage, transparency, accountability, reliability and independence from the big institutions, who might otherwise control your data and charge you fees. But technology has moved on, rendering the blockchain design pattern, which was and remains fundamentally flowed, redundant. A good analogy is the laser disc. It arose out of the need for something better than magnetic tape yet in 2020, it has completely disappeared; replaced with cloud storage and streaming. Blockchain was first proposed some 12 years ago when massive computer centres, owned by big corporations, controlled everything. Data sovereignty, security and cost all seemed to go per favour of the big corporates. Blockchain promised to save us.
In the decade since, it is curious that the only commercial application for blockchain is for bitcoin and crypto-currencies, replete with public disasters, mismanagement and all the latent criminal involvement. It is also no surprise that bitcoin is the darling of Slavic and Russian ransomware hackers who use blockchain to achieve anonymity in order to perpetrate their extortions. This is the exact antithesis of what blockchain was supposed to be all about.
However, I do not want to concentrate on blockchain implementations or judge them. What I want to do is analyse what blockchain was supposed to be yet isn’t, the problems it was supposed to solve but didn’t and why it is damaging to the international datasphere. My company has spent 40+ years writing centralised and distributed data transaction and storage systems for big retailers, large insurance companies and State courts. We employ the latest tech from the largest corporations and build private clouds for those who insist. We live on the bleeding edge of both traditional data and IoT systems. We eat and breathe technology yet most blockchain spinmeisters haven’t got the first clue about the technology they spruik and are never available to debate its merit. I want to look at the underlying technology and assess blockchain for relevance in 2020.
1. What is Blockchain
It is important to agree on what blockchain is, even if we disagree on the ‘why’. Blockchain is basically an encrypted file system that is spread across multiple servers at multiple geographical locations, all linked with highly encrypted indexes. It included multi-layer redundancy and an encrypted management layer to execute and control access. The idea was that we should be able to disconnect ourselves from these server behemoths and utilise lots of smaller, independent servers. No one site, institution or government would be able to interfere or technically, impose fees or control access. Further, each of the blocks would keep a complete history of changes – by whom and by what – thus making for detailed audit trails, guaranteeing the authenticity of each transaction. The goal of blockchain was thus quite utopian but it was hijacked early on by crypto currency implementations, driven by the need for secrecy and anonymity. So what does blockchain do better than access methodologies? Frankly, nothing.
2. Execution Cost and Server Complexity
The first thing to note is that all this encryption and distributed nature incurs enormous computing overhead. Accessing even one byte of data now becomes quite a computational exercise because it needs to be found from an encrypted index, an end point (or multiple end points) computed, the data decrypted and finally retrieved. A dizzy process at best. Contrast this to a regular data store: there is a single endpoint that his known in advance. It turns out that servers to execute all this encryption and indexing are very expensive and very large. In other words, those large corporate data centres that blockchain was intended to unshackle us from are the only genuine sources of such computing power to deliver blockchain of any scale. Conclusion: compared with traditional transaction technology, blockchain is overly complex and expensive.
3. Latency and Network Performance
The rise and rise of large public clouds (Microsoft, Amazon, Google, IBM and a host of others) and even the popup suburban data centres, have been made possible by very low cost storage and very high bandwidth networks. What we have all found is that it is important to collocate not only application data into one location but also its attendant applications. This is required in order to reduce access latency, a critical component to application performance. We will come to redundancy later but this geo-collocation is important for high performance. Blockchain represents two significant detractions: speed and cost. First, it’s distributed nature makes it incredibly slow by comparison. Anything computationally data intensive will be excruciatingly slow but under blockchain’s distributed data model, all this access complexity takes a significant toll. So slow in fact that it will not be tolerated. Secondly, it loads down the internet with useless retrieval traffic. Simple data calls that would normally require one DNS lookup now might generate 100’s. The traffic is splintered and must by its nature come from multiple (again, even 100’s) locations. This is because every data call must go to multiple locations to reconstruct its queues or ledgers. In consequence, the internet suffers massively with marshalling and indexing traffic for no useful purpose other than to retrieve the required chain. Any blockchain application should be made to pay for this additional traffic because it reduces overall bandwidth availability. Conclusion: blockchain is a disruptive drain on international computing capacity and too slow for most business implementations.
10 years ago, one of the big risks to data storage was the proverbial jumbo jet crashing into your data centre. Even power outages, hurricanes and other natural events could cause massive disruption to data availability. Blockchain was seen as a way to solve this because it had inherent geolocational redundancy. Mind you, this assumed that your dispersal was wide enough to compensate. Wind the clock forward just 5 years and georedundancy became a standard feature of all public clouds. In 2020, with extremely low storage costs, georedundancy is effectively built-in to most cloud storage solutions so this is no longer a problem. Conclusion: blockchain isn’t solving a redundancy or storage problem we currently have.
10 years ago, data security was a problem and data theft and corruption was commonplace. Blockchain, with its encrypted distribution model and encrypted data blocks was seen as a way to combat the cybercriminal. Apart from some notable disasters, cryptocurrencies have largely achieved that but only after dispatching any semblance of transparency. Fast forward to 2020 and almost all data stores encrypt their traffic, offer database, row and data element seamless encryption. Conclusion: we simply don’t have the security problem that blockchain might have resolved.
6. Location of Data
The European Union and the USA have strict rules for corporate entities as to where their data is housed. In the USA case, it prohibits offshore storage. Under a blockchain model, you don’t know where the data is stored, nor are you allowed to know. What if it is all housed in Russia or China for a US corporation? Conclusion: blockchain is problematic for data security regulations.
7. Audit Trails
One of the facets of blockchain that interested me early on was the concept of an implicit audit trail. Every change was recorded: when and by whom. But in reality, all data stores already implemented trigger technology that can achieve that – we just didn’t use it. That technology was available 20 years ago. But back then, we all realised that it doubled or even quadrupled the necessary data storage cost and dramatically slowed down performance. That is still true today. We also realised that probably 98% of data collected does not change or is relatively archival. It was seen as pointless and expensive placing deliberate audit trails over static data. Perhaps we got lazy in some instances but by and large, we didn’t want or need the enforced audit trail blockchain imposed. We would choose when we needed it. Conclusion: we already have control over data auditing so blockchain doesn’t help. In fact, it is a net negative because it enforces it.
My last issue with blockchain is harder to monetise without giving hysterical figures. One of the strongest sets of brakes on application development has been the money already invested in applications. Once you’ve spend millions on your applications, any corporation needs to generate a return on that investment. Further, the scale of these applications make it very difficult to just shift the underlying data access technology. Why else do we still need COBOL programmers? In the last 4-5 years, most major software vendors have given us the ability to integrate, migrate and distribute to different platforms but all of these are based on traditional data access models. To turn around and implement or adapt legacy applications to meet the blockchain methodology will be a cost most institutions will not want to bear. And if it doesn’t help achieve anything new and useful, what would be the point? Effectively, blockchain applications such as Bitcoin, needed to evolve for and with blockchain. It is telling that in 10 years you can hardly point to anything outside of the cryptocurrency world that employs blockchain. Conclusion: blockchain is not going to achieve momentum because of the high application development cost required.
My conclusion is that the blockchain concept belongs on the historical scrap heap, right alongside cassette tapes, floppy disks and CD ROMs. It promises nothing we cannot already achieve natively, with far less cost and complexity. Blockchain is an inordinately expensive alternate and a poor network citizen so it is time to stop talking about and investing in it. We just don’t need or want blockchain.