I will start right off with my conclusion, to get your attention. If done correctly and optimally, EOS only needs to pay 1% inflation to BPs, and the rest of this post shows my justification. And this is coming from someone who would love to be a chosen BP.
Second, I will state my disclaimers. By no means am I a blockchain expert or guru, so I don’t know if my thoughts in this post will be met with ridicule or seen as enlightening. The post is just one person’s opinion(s). However, on this subject matter, I do think my opinions are relatively qualified, since I have been the owner/operator of an IT networking company for the last 20 years. My company, Free Space, has designed, installed, tested and optimized over 2500 enterprise networks. My second disclaimer is that I own 1000 EOS coins, which is about 25% of my crypto portfolio. This is what started my research, as I wanted better to understand the BP ecosystem, to best protect my investment.
When I compare the decentralized design of the EOS block producer node to most of today’s enterprise networks, I see very few challenges. From a size perspective, the EOS network will only be a medium-sized network compared to global enterprises that can have thousands of locations, each larger than a blockchain node. Additionally, there is not much complexity, as node design is extremely straight forward. Each node is not much more than a small LAN, server farm, and secure WAN connectivity. Of course, the blockchain software and its upkeep require skill and expertise, but from a pure networking perspective, things are rather straight forward.
To me, the intriguing part of blockchain infrastructure is the decentralization. However, this is a double-edge sword. We all know the advantages of decentralization which include things like security, reliability, uptime, and resiliency. However, since each node is designed and operated independently, it brings some challenges surrounding consistent performance, network monitoring, scaling, accountability, etc.
With that said, my mind immediately steered towards a few critical questions. How can we best recognize the advantages of decentralization, and at the same time, minimize the disadvantages? And how can we do it in the most economical way, without jeopardizing the integrity of the blockchain? After much thought, I came up with the following concepts.
First, I think there should be a tiered system. Tier 3 would be the highest and would be the primary nodes that run the blockchain. There should be up to 50 of these nodes. And any give time, 21 of these nodes will be the block producing nodes. The remaining need to be equal in every way, as they could be called on to produce blocks at any time. Tier 3 nodes must be fully redundant with electric, servers, bandwidth, engineers, etc.
Tier 2 would be funded less, but still fully capable of producing blocks. However, they may not have the redundancy and engineering support like Tier 3 nodes. These nodes will be critical to provide emergency support, testing, and training. There could easily be a hundred Tier 2 nodes.
The final Tier 1 nodes would be for the enthusiast. These would be single server nodes hosted at secure data centers. In all reality, these nodes would not have redundancy, would be owner/operated and will never produce blocks. But they are very inexpensive and will provide invaluable performance metrics, that may be used to optimize Tier 3 nodes in the future. Additionally, tier 1 nodes would act as incubators for engineers and blockchain programmers to insure the future of the EOS ecosystem (we want them working on EOS and not the competition). There could be hundreds of Tier 1 node operators. Finally, there will always be the hobbyist, who wants to run on pcs and local internet connections. The community should support these individuals but should not fund as they would not qualify at tier 1.
A second critical topic is performance evaluations and monitoring. I have wrapped this solution in with the solution to another topic, which is proper testing and roll-outs. I would suggest that we have an official test network. These have proven invaluable in the lead up to the mainet roll-out. I participate in the Cryptolions JungleTestnet, and only have wonderful things to say about that experience. With the counts in my tiered system, we could have a program, where every node spends one month a year on the test network. And on this network, we could do performance, security, and vulnerability tests on each node since they will all use different hardware, connectivity, configs, etc. This would insure Tier 3 nodes were adequately prepared for growth, and that the Tier 2 nodes can step up if needed. It would also provide great data about each BP’s performance to EOS coin holders, so they can cast their votes with accuracy. The test net would also be used to test new software versions, new dApps, new monitoring tools, without putting the mainet at unnecessary risk.
Now, lets get down to the brass tax. What will all this cost? First, we need to understand the mindset of a BP. BPs are going to invest cash and time, and not receive any fiat compensation. Many will line up to take this risk. But I don’t want any average joe running the network that I have staked. While there are many critical components that will affect the success of EOS, without question, if the nodes don’t perform, our investment does not perform. So, we need to attract the best, realizing that great network engineers, linux engineers and blockchain coders have no shortage of opportunities. So, we must make sure our funding is solid, to attract these skills to the EOS ecosystem. But on the flip side, if we overpay for these skills, then profits and growth could be limited. Entrepreneurs look for 20-30% return on their investment. However, on elevated risk ventures, they need higher returns. Furthermore, if the returns are not in fiat currency but coins, whose value can fluctuate greatly, that return needs be even higher. Personally, I think 2x is just about right. For example, if a block producer has 100k in out-of-pocket annual investment, they should be compensated with 200k of coin in present value.
Ok, so now we can dig into some numbers. For illustration purposes, my estimates on costs are purposely high.
Tier 3 node- annual costs
Data Center facilities - $75,000 year
Server support, community projects, Dapps Engineering - $500 k year
Accounting, legal, HR, Operational costs - $50,000
Marketing, outreach, events - $75,000
Annual equipment / software replenishes - $50,000
Total = 750k annual
So, a Tier 3 node can be operated profitably for 1.5M in EOS coins. And, in my opinion, any funding over 2M for operations of a single node, is wasted funds. Keep in mind that many Tier 3 BP will expand operations for consulting, coding, and dApp development to third parties and thus have higher out-of-pockets costs. But said 3rd parties should pay for these services, and thus they should not be funded by EOS coin holders. EOS coin funding should be based on the costs to run the node and what the BP invests back into the community.
Tier 2 node - annual costs
Data Center facilities - $24,000 year
Server support, community projects, Dapps Engineering - $150k year
Operational costs - $20,000
Marketing, outreach, events - $20,000
Annual equipment replenishes - $20,000
Total = 235k annual
So, a Tier 2 node can be operated for 470k in EOS coins. And any funding over 500k for operations of a single tier 2 node, is wasted funds.
Tier 1 node - annual costs
Data Center facilities - $2,000 year
Server support, community projects, Dapps Engineering - $10k year
Annual equipment replenishes - $5,000
Total = 17k annual
For 25k per year, we can keep hundreds of enthusiasts investing their time and money in the EOS ecosystem. That is 250 future geniuses managing nodes for EOS and not the competition.
So, the total network cost look like this:
50 Tier 3 @ 2,000,000 = 100,000,000
100 Tier 2 @ 500,000 = 50,000,000
250 Tier 1 @ 25,000 = 6,250,000
Total annual cost = 156,250,00
Annual inflation = 0.927%
So, in conclusion. I am confident, that should EOS coin holders want to vote this way, that we could have a robust tiered BP system. It could have 150 nodes capable of producing blocks and hundreds more in support. We could have a robust testing environment with network wide performance standards. And we can do it all for 1% inflation. And this is based today’s market cap of 16B. As market cap goes up, so will the requirements of each node. But surely, the market cap will increase faster than additional expense requirements. So, that would imply that the starting rate of 1% could go down as EOS market cap grows.
I hope this information was found valuable and helps the EOS community finalize how we want to pay out block rewards and which BPs we want to support and with how much funding.
Free Space is an official BP candidate and would love to see the community adopt a tiered approach as I have outlined and would love to be voted in at the tier 2 level.
https://bp.eosgo.io/listing/free-space-block-producer/
https://steemit.com/eosis/@pythagoras345/free-space-eos-block-producer-candidate
https://t.me/pythagoras345
https://steemit.com/@pythagoras345