Qitmeer—Medina Network Operation Report

Qitmeer Network is a BlockDAG network based on the PoW consensus algorithm . It is an exploration of the integration of the classic UTXO model and cutting-edge DAG technology , and adopts the GPU graphics card mining mode. Since the launch of Medina Network at 16:00 on December 30, 2019, as of February 18, 2020, the Qitmeer test network has been running for 50 days.

Medina Newtork is guided by the principles of openness, equality and inclusiveness . Anyone can participate in mining and network node establishment and become a member of the Qitmeer public test network.

Regarding the operation of Medina Newtork since its launch, in the following, we will count and display parameters from multiple aspects such as the number of blocks, block time, coin production, full node status, concurrency of DAG network structure, and computing power , and comprehensively observe the overall operation of Medina Newtork.

1. Introduction to mining principles

In terms of protocol, Qitmeer adopts the BlockDAG collaboration model, which is fairer than the competition model of blockchain. In order to balance the relationship between CPU/GPU mining and ASIC mining, and avoid the project (at least in the early stages) being monopolized by mining tyrants, the choice of mining algorithm follows the principle of "one machine, one vote" of Satoshi Nakamoto , and introduces the Cuckoo Cycle proof-of-work algorithm, which has anti-ASIC characteristics. The Cuckoo Cycle proof-of-work algorithm is a memory-hard algorithm that is used to find certain subgraphs in large pseudo-random graphs. Compared with the SHA256 algorithm used by Bitcoin, its calculation process requires higher memory bandwidth than the operation frequency.

Medina Newtork (Medina period) starts with a low-entry-threshold pure small-image mode (edge_bits=24). It will last no longer than 365 days, and will transition to a higher-entry-threshold, more efficient GPU utilization large-image mode (edge_bits=29).

To participate in Qitmeer mining, you can currently use the official software qitmeer-miner , which supports Windows, MacOS, and Linux operating systems. It supports GPU mining . NVIDIA and AMD GPUs require the installation of CUDA and OpenCL plug-ins.

During the testnet, only cuckaroo is currently supported. In small image mode (edge_bits=24), at least 1GB of video memory is required. After switching to large image mode (edge_bits=29), a graphics card with at least 6GB of video memory is recommended.

2. Current network operation status

As of the early morning of February 20, 2020, approximately 156,000 blocks have been generated, and approximately 18.8 million PMEER coins have been produced, accounting for approximately 9.4% of the total coins in the test network.

The block time of the 156,000th block (order) is 2020-02-20T05:53:40+08:00 (corresponding to the timestamp 1582149220), and the corresponding main chain height (height) is 148864. The block time of the first block (order) is 2019-12-30T15:55:42+08:00 (corresponding to the timestamp 1577692542). The overall time span of 148,864 main chain blocks is 4456678 seconds, and the average block time is 29.94 seconds, which is consistent with the set value of 30 seconds. If calculated based on 156,000 order blocks, the average block time is 28.57 seconds, a difference of 4.77%.

When Order=156000, the main chain height is 148864, and the overall expansion rate of the DAG network is f=(156000-148864)/148864*100%=4.79%.

If we regard every 200 blocks as an instantaneous state, and define its instantaneous concurrency rate f=(order increment-height increment)/height increment=(∆O-∆H)/∆H, and count the network concurrency once every 200 blocks (order), we can get the following concurrency rate curve:

Overall, the concurrency rate is stable at less than 10%, and in most cases it is around 5%, without too many concurrency surges.

The following figure shows the difficulty change trend curve from block height (order) 1 to 156000 (corresponding time is 16:00 on December 30, 2019 to 5:53 on February 20, 2020):

This figure is a difficulty change curve obtained by counting the block difficulty every 40 blocks. It can be seen from observation that the difficulty value within these 156,000 blocks shows the following changes: the difficulty change is mainly stable in the difficulty band between 6,000,000 and 10,000,000, fluctuating around the 8,000,000 difficulty line, with an upper and lower fluctuation range of about 25%; the highest difficulty increase occurs between blocks 30,000 and 45,000, and lasts for about 5 days, which should be the change brought about by the influx of new computing power, and then falls back to near the 8,000,000 difficulty line.

Since the difficulty changes with the computing power, there is a certain correspondence between the difficulty and the computing power: hashrate=diff/scale*50/30. In the small picture mode with edge_bits=24, the scale value is 48. By conversion, the minimum computing power level required under the current difficulty can be obtained, and thus the computing power change curve reflecting the computing power level of the entire network can be obtained:

As can be seen from the above figure, the stable computing power level of the entire network is about 200-350kGPS. The computing power of a GTX1080Ti is about 65GPS, and the computing power level of the entire network is approximately equivalent to 3077-5385 GTX1080Ti.

The total network computing power level obtained by the above method is the minimum computing power level required by the current difficulty. It may lag behind the actual computing power level and may even be lower than the actual computing power level, but the difference is not large. Therefore, it can effectively reflect the computing power level of the entire network.

The power consumption of a GTX1080Ti is about 5 kWh per day, and Medina Network can generate at least 2880 blocks per day, which is 345,600 PMEER. According to the computing power level of the whole network of 200-350kGPS (3077-5385 GTX1080Ti), the daily electricity consumption is about 6923-12116 yuan (the electricity price is calculated at 0.45), and the corresponding average electricity cost of a single PMEER is 0.02-0.035 yuan. Due to the concurrency of the DAG network, the actual daily coin production will be higher than 345,600, and the concurrent coin production will be about 5%, so the actual electricity cost will be 5% lower.

According to the current computing power level on February 20, it is about 300kGPS, and its average electricity cost is about 0.03 yuan.

The current price of PMEER is about 0.22 yuan. Excluding electricity costs, there is still a large profit margin.

According to long-term observations, the number of Medina Network's long-term stable online full nodes is about 40-50. The current number of connections monitored on February 20 is 39 nodes.

Since the current computing power is basically concentrated in the two major mining pools, meerpool.com and 666pool.cn, there are few people who set up full nodes for independent mining. As a network that has just been running for 52 days, the current number of full nodes is reasonable. However, in the long run, in order to grow the network, the expansion of full nodes should be given top priority.

3. Conclusion

Qitmeer's public test network, Medina Network, has been running relatively stable for 50 days since its launch. The overall stable computing power level is around 300kGPS, and the number of long-term stable full nodes is at least 40. The current block output has exceeded 156,000 blocks, and the coin output has exceeded 18.8 million PMEER, accounting for about 9.4% of the total coin output of the test network. The overall expansion rate of the current DAG network is about 4.79%, and the average block time of the main chain is 29.94 seconds, which is consistent with the set value of 30 seconds.

The Qitmeer network, which combines the classic UTXO model and cutting-edge DAG technology, is moving forward steadily.

Qitmeer Economic Model Group