Meet Magister Ludi: The Process By Which Validators Are Chosen on the Free TON Blockchain

Free TON is a distributed system that uses the Proof-of-Stake algorithm of the BFT consensus to work consistently.

The efficiency of the algorithm directly depends on the number and quality of network validators. Therefore, after the initial launch of the blockchain, the Free TON community set a task to attract competent, motivated validators and at the same time prevent the concentration of resources in one place.

The required number of validators was determined to be several hundred. This goal is a real challenge, and in the case of Free TON, it was called Magister Ludi.

The Crystal Game contest was announced in June 2020 and lasted for more than six months. Its winners were awarded the title of Magister Ludi — "Master of the Game". Another result is a set of 375 new Free TON blockchain validators. We talked to the organizers and winners of the contest about its main results, difficulties and achievements.

"Capture The Flag" For Validators

Although this is called a game, the validators had to solve non-trivial technical problems, respond to emergency situations and identify possible errors in the node software.

Initially it was planned that the game will last four weeks, during each of which they will carry out nine validation cycles with a gradual increase in network load: from a routine check in the first stage to network overload at the last. For the performance criteria of the validator, special metrics have been developed that take into account the percentage of blocks produced, delay times and much more. 

Magister Ludi lasted longer, adapting to unexpected challenges and changing the reward system. The period for which the prize money was awarded and the number of winners changed: instead of 300, there were 375 of them.

The winners of the game, based on the results of four stages, received a lock-stake award for a period of two years in the amount of 375,000 to 750,000 crystals, depending on their place in the ranking. Now validators can use these tokens for validation, which will bring them their own income. These funds allow the winners of the contest to keep the network in their hands and make it more difficult to enter the validation of teams whose competencies have not been tested “in battle". In addition, validators receive a vesting reward — from 120,000 to 240,000 crystals.

Organizers Sum Up The Results Of Magister Ludi

Results And Achievements

Talking about the goals set, Sergei Zaitseff, Senior Data Scientist TON labs and Product Owner of TON Live Blockchain explorer, noted that the game solved the following main tasks: recruitment of validators who will gain the necessary experience in working with the blockchain during the contest and testing various network parameters.

“I think the contest was great. Difficult, but interesting. We have completed all the technical tasks that we set for ourselves.” Mitja Goroshevsky, CTO of TON Labs

More than 400 participants applied for the contest, 379 participating teams passed the cut-off threshold for points scored, and 375 teams passed the KYC procedure. At the moment, about 370 participants have received the award, who were subsidized in order to be engaged in network validation in the next two years.

As a result of the contest in December 2020, partial decentralization was announced — now Free TON has not 40 validators, but 400.

Sergei explained that at the moment this is more than any PoS blockchain and more than enough for Free TON, so there is no need for new validator recruitment contests yet. A corpus of validators has already been formed, which, in his opinion, is well versed in validation issues. In their Telegram channels, the validators answer the questions asked and explain the necessary things to beginners.

In addition to recruitment of validators, the game helped to evaluate the parameters of the network and draw conclusions about what needs to be improved in the protocol design.

“First of all, we will deal with security issues and improve the work of external messages. This part of the architecture will have to be revised almost completely." Mitja Goroshevsky

Decentralization And Geography Of Validators

Sergei Zaitseff said that initially it was announced that one team will receive only one prize: 

“Our goal was to ensure that a large number of crystals did not fall into one hand, and there was no monopoly. The KYC procedure was used, and although it was quite soft, and someone may have managed to get 2-3 prizes per team, but no more than that. There was no concentration of prizes in one hand." 

Sergei also commented on the discussion about a greater degree of centralization due to the participation of the Korean community, noting, in particular, that a significant number of Korean participants and winners, and there are about 160 of them, is due to the fact that they used hired DevOps, who themselves participated and helped other participants.

According to Sergei, at the moment, approximately 40% of validators are the Korean community, and 60% are the rest of the world. Chinese validators are beginning to participate more actively, and participants from Western Europe appear — Italians, Maltese. After the merger of the Dune Network with Free TON, the French community will also be added.

A Word To The Free TON Validators

Anatoly Ustinov, one of the Free TON validators and contest participant, said that he came to the community at the very beginning of the game of validators. He knew about the TON project of the Durov brothers for a long time. He liked the technology itself, but their financial model did not suit. One of Free TON community developers convinced him to join, who once spoke about Free TON and the principle of meritocracy — tokens in the hands of the capable.

To participate in the contest, Anatoly used AMD Ryzen 9/128GB RAM hardware, for 100 euros per month. The nodes were managed by himself, so there were no more operating costs. Anatoly believes that there were no difficulties with participation, everything was quite simple and understandable.

Validator turb0cat said that he entered the game at the very first stage in May 2020:

“I followed the fate of the Telegram TON and when the news about its closure appeared, I carefully looked at who and what was doing on its basis. The guys from TON Labs won over with their openness and approach to business."

The validator said that he used a 12-core machine (24 threads) with 128 GB of memory and SSD disks to participate in the game. Also experimented with a 32-core machine and 512 GB of memory, but this did not lead to a significant improvement in the results. Renting a server cost an average of 100 euros per month, and at the time of the experiments up to 300 euros. The time spent is from 8 to 40 hours a week. In his opinion, a lot was decided by the location of the server and the analysis of competitors — the closer the server is to the network core, the better the results. About the process itself, the validator responds as follows: 

"There were some unpleasant moments with the scripts failing, which led to the loss of place in the ranking, but that's what the game is for." 

"Whoever wants to be a validator becomes one. Validation is not for the vulnerable and impressionable, you need to be understanding. In difficult and emergency situations, no one will personally hold your hand. This must be understood." Anatoly Ustinov

Previously published at https://freeton.house/en/magister-ludi-results-of-the-first-large-scale-free-ton-validator-contest/