Two scholars affiliated with the Cambridge Centre for Alternative Finance, in Great Britain, have prepared a fascinating overview of the present state of blockchain and distributed ledger technology (DLT). Dr. Garrick Hileman and Michel Rauchs, respectively a Research Fellow and a research assistant, have gathered data from more than 200 private and public sector organizations from 49 different countries around the world to benchmark “the emergence and evolution of the DLT ecosystem.” Layers on Top of Layers Hileman and Rauchs make the following points:

  • The protocol layer of this ecosystem is only slowly maturing, and the limit of this maturation is one of the key challenges for the broader adoption of DLT;
  • Most users experiment with only small-scale, isolated layers, with live applications allowed only as “permissioned” layers;
  • There is an increasing focus on creating common standards of enterprise DLT frameworks that will allow for interoperability, but this effort to overcome fragmentation is itself fragmented into “a variety of consortia”;
  • Ethereum in particular has been tested at 57% of central banks, either via the public network or a permissioned version.

A “protocol layer” consists of the core software elements that make up a distributed ledger. It is distinct from both the network layer built on top of it, that is, the actual peer-to-peer stuff, and from the application network built on top of the network. A “permissioned” layer or network s what the adjective suggests, an arrangement in which only selected parties can make changes to the distributed ledger. These authors maintain that many publications have hyped the significance of bitcoins, focusing only on the positive, not the drawbacks. The hypsters say for example that blockchains require no trust among counterparties. That is false. These authors respond that blockchains “always require some degree of trust.” Likewise, hypsters regard blockchains as ideally secure. Hileman and Rauchs reply that they are “not more secure than other systems.” Attackers could in principle acquire the private keys of some network participants and thereby get “full access to the shared database, including the ability to reverse transaction history.” About that Ecosystem Hileman and Rauchs estimate that enterprise DLT start-ups now employ approximately 2,000 people full time. Beyond that, established corporations have several thousand employees focusing their efforts on DLT based work. The most labor intensive element of the system consists of the infrastructure providers, who either develop the protocols or build the networks, or both. These firms have “twice the median number of full-time employees as [do] application developers and operators.” In simpler terms, the median number for those three types of employers is 10, 5, and 5, respectively. Nearly half of the surveyed DLT start-ups are in North America (47%). The second largest group by continent is that which hails from Europe (28%), then 19% (Asia Pacific). But employment isn’t precisely aligned with the number of enterprises on each continent. A full 61% of employees working in this field are in North America, and only 13% are in Europe. On a more technical issue: no app providers fully open source their codebase. But 8% of network operators do just that, as do 27% of infrastructure providers. More impressively perhaps, one third of the infrastructure providers that have not already opened up their codebases say they plan to do so in the near future. Infrastructure providers have a wide range of user types. A majority of infrastructure providers say that their main customers come from the financial sector (which includes both banks and FinTech companies). But their other users can be quite diverse, whereas operators tend to specialize in a way that gives them a quite specific clientele. More than three quarters (78%) of operators have four or fewer user types. That is true of only 29% of infrastructure providers. Architectures The paper also tells us that the architectures of DLT systems “have significantly changed since the first blockchain implementations, which were largely based on Bitcoin’s original design.” One of the trends since then, though, has been toward “limiting the data that needs to be stored on the network itself for various reasons, including privacy concerns, data storage constraints, processing costs, and network latency issues.” In some cases the blockchain only stores fingerprints of the data in the form of hashtags that reference an external source.