SONM — A Fog Computing Platform
TL;DR. Ecology:3. Technology:2. Decentralization:2. Valuation:4. Rating:3/10
(Update 2018-12-15: Please see the semi-annual ratings review of SONM here.)
SONM (Supercomputer Organized by Network Mining)was introduced 24th of January 2017 on bitcointalk.org by Andrey Voronkov, then the owner of DrugDiscovery@home project. The stated goal was to apply widely used BOINC technology together with Ethereum smart contracts to create a distributed general purpose supercomputer where buyers of computing power could get a better price than when buying from centralized solutions from Amazon, Microsoft or Google.
After teaming up with a couple of prominent Polish members of the DrugDiscovery@home project, there was a fallout resulting in these Polish team members quickly leaving the SONM project while warning everyone it was a scam. These accusations were met by a rebuttal by Andrey and it seems the infighting came from different opinions about whether SONM was to utilize BOINC technology or not. Andrey was of the opinion that the project ought to stay away from BOINC integration. In this same rebuttal, he transferred the lead role of DrugDiscovery@home from him to the two Polish team members, effectively distancing himself from that project. More confusingly, he seems to have stepped down as lead developer of SONM at the same time, to instead become just an advisor. Sergey Ponomarev, also a co-founder, took the new role as SONM lead developer. As BOINC was abandoned, the project instead decided to utilize the Cocaine client (open source Platform as a Service). A couple of Cocaine developers joined the SONM team at the same time. By using Fog computing (a distributed form of cloud computing), SONM aspired to connect suppliers and buyers of computing power in a decentralized way. Suppliers will earn SNM tokens that are provided by the buyers; this makes the utility token classification quite clear.
A pre-ICO was scheduled to run for 10 days starting on the 15th of April 2017. The max limit of 10 000 ETH raises was achieved within a day however. Immediately afterwards, an extensive bounty campaign was initiated. Two months later, on the 15th of June, a second, larger ICO was held. It was sold out in 3 days and raised USD 42M. The SNM token is built on Ethereum and is of the ERC20 standard. It was initially called Dilatones (DIL) but that has since changed. BD Ratings did not find details on the amount of tokens distributed to the team as the initial whitepaper is no longer accessible on the SONM webpage.
On the 24th of December 2017, SONM released a MVP so that users could try it all out on a testnet. A more advanced version was released on 30th of May for Rinkeby testnet. Information was also given that it would run for a month after which it would be published on mainnet where the real market activity awaited. Finally, due to temporary congestion on the Ethereum network, it seems the team is planning to set up its own blockchain with a Proof of Authority consensus algorithm. Details on this blockchain is quite scarce however, and BD Ratings has found no information on the issuance model (if any) or if the SNM tokens decouples fully from the Ethereum chain on which they are issued today.
The SONM project has made use of multiple bounty projects, where mainly bitcointalk.org users are paid to change avatars, change signatures, retweet posts on Twitter, write (positive) blog posts etc. The only advantage of this, and it is debatable, is a broader distribution of tokens, but on the whole these bounty campaigns are hurtful for cryptocurrencies and the space. They encourage the superficial instead of the underlying code, short sighted "pump and dump" instead of long term sound money, illoyal bounty hunters instead of loyal users of the token.
BD Ratings also questions the choice by the SONM team to run two separate ICO's, the latter planned for just a month after the first one (it was delayed another month). The first ICO was termed 'pre-ICO', but that is just semantics. They did in fact have two ICO's, to maximize the amount of raised money. It indicates yet again a focus on quick riches instead of actually building for the long term. This price focus is seen again when the SONM team encourages people to buy tokens. And as SONM pays for articles. They also seem to pay for YouTube reviews.
During the first ICO a well-known escrow, SebastianJu, was used to add legitimacy and security to the ICO process. This user has built up a service where he provides himself as an intermediary during ICO's, so that no-one easily can run away with all the funds. Apparently, the SONM team disregarded his services as well as his attempts to get clarity of the situation as the second ICO were to start. SebastianJu was ignored, and was finally told his services were no longer needed as the ICO contract was a multi-sig. The multi-sig was controlled by SONM team members as well as an advisor. The whole thing looks very unprofessional.
And why the low level of professionalism? It might have to do with the fact that the SONM founders seems to have contracted a really shady ICO promotion company aptly called ICOPromo. Reviewing their webpage, you will see that other clients are scams like TAAS (promises extremely high yield) and Paragon. So organized scammers are in charge of SONM's ICO activities with complementary bounty programs, that's great.
Despite all the above, the SONM ICO sold out quickly, raising USD 42M. This speaks volumes about the complete craziness of the cryptocurrency ecosystem at this time. Remember, SNM tokens are not on their own chain; during the ICO it was fully transparent that they were in fact utility tokens built on top of Ethereum. People bought without any clue how to value these utility tokens.
But BD Ratings wants to add on a last note that a strong ecology might indeed grow out from what the SONM team is building. They are the experts in distributed computing, not BD Ratings, and while they have blundered in the many aspects described above, a good product might very well sweep those facts away with force.
Reasons: Mainnet not yet launched. Myopic focus on marketing and new money.
In this Technology section BD Ratings will focus on what consensus algorithm will be used for the blockchain and not whether SONM uses the proper computing model. SONM has chosen not Cloud or Grid computing, but Fog computing. Whether this is prudent or not may be established later as we get data on how much the utility token is used to buy computing services. There are indications that SONM is indeed very strong in this area, and they have hired proven specialist as well. This analysis however focuses only on the money-like properties of the SONM token, which can be estimated without digging too deep into the different computing models that the SONM team seems to be specialists on. For a valuation of a utility token, a ratings agency has to really understand the underlying market. This makes the task overwhelming as there are as more utility tokens in the cryptocurrency space than there are industries. Valuing a token in the sense of its ability to storing value, like Bitcoin, other tools can be used.
In the end of January 2018, it was revealed that SONM had hired the former team lead developer of Yandex Cocaine, the platform on which SONM builds the computing power marketplace. Although probably great in terms of product development, BD Ratings has found no information on cryptographers being added to the team. This distinction is important; the potential increase of utility of SNM tokens matters in the long term only of the whole blockchain is secure and decentralized.
More information on the Proof of Authority blockchain under construction is needed to set a proper Technology grade. If there is indeed no issuance of new SNM tokens for each block, transaction fees ought to be rather high to incentivize the master nodes to stay online, which begs the question why SONM moved away from ETH in the first place. Having mainly master nodes run by SONM is not enough as it mitigates the logic of having a blockchain altogether. The SONM team has provided some information on the future master nodes, with ROI calculation, but has also not yet described in detail where the payment comes from. BD Ratings will update this rating as soon as a clear picture of the new blockchain emerges.
Taking a quick look at the development of the Fog computing client, it seems good progress is ongoing and there are multiple code contributors as well. There doesn't seem to be a clear repository for the blockchain development yet.
Reasons: Mainnet not yet launched. Distributed computing experts on the team. Few if any cryptographers or blockchain specialists on the team. Few details on the PoA algorithm.
SONM is a company. So far, very little decentralization exists within the project. There is an outspoken vision of future decentralization, but even the new Proof of Authority algorithm will be setup in a way that initially gives SONM full control of the ledger. They say that it will be decentralized in the future, with a regular PoS structure instead, but to take that as fact does indeed require trust. The SONM team does seem to understand the need for decentralization but they need to apply this practically as well before a higher Decentralization grade can be set.
Why does it matter that SONM is a company? This is a central point of failure right now, and if something were to happen with this company, the SONM blockchain might be in danger. To establish an anti-fragile blockchain, SONM needs to distance itself from being the one and only pillar of the project, and luckily this is their stated plan as well.
The code is public on GitHub, and there are no stupid patenting practices going on within the project.
Reasons: Company with control of the ledger.
The 62M valuation of the SNM tokens is nothing extreme when compared to other native blockchain tokens, but it is still quite high for something that is not yet live on mainnet with an own, decentralized blockchain. The main focus for SONM has been to view SNM as a utility token, with focus on the fog computing technology and not on blockchain building. With a planned native blockchain, this will change and the tokens can then be easier to value in relation to other store-of-value coins. That however requires extreme care as managing an own blockchain instead of just relying on ETH demands real expertise.
Reasons: No mainnet. Still centralized. No extreme valuation when compared to other coins.