A Few Thoughts on DigiSpeed & Blockchain Scalability
Question #1: Why does DigiByte need to be able to match Visa's average transactions per second?
Short answer: Utility, usability, marketability and growth well into the future. "DigiSpeed" has already became a buzzword in the crypto community and has a very marketable appeal moving forward. More importantly DigiByte takes a bold step forward with DigiSpeed to dramatically scale a blockchain with multi-algorithim mining. However, this does not mean the DigiByte blockchain size will jump dramatically over night.
Question #2: I can't afford to mine and run 8 MB blocks on my PC?
Short Answer: We are a long way off from hitting 8 MB blocks. We are not even coming close to filling 1 MB blocks. This change simply removes a cap that Satoshi himself even admitted was an arbitrary limit he put together on the fly. Our assumption is we will not see 8 MB blocks until several years from now.
Question #3: What incentive will people have in the future to run a full DigiByte node when there is a large amount of data in the blockchain?
Short answer: The "DigiSize v5.0 wallet we are planning for mid-late 2016 will tackle this issue head on. This will include advanced pruning and syncing features as well as some increased incentives for running a full node. We are tossing around several ideas for what this incentive might look at. Perhaps a staking feature, perhaps a new type of mining algorithim or perhaps some other feature yet to be thought up. We would like all the input we can get on this topic. In the end the key to long term survival and network security is tens of thousands of people around the world running full DigiByte nodes.
Question #4: Won't people attack the chain with bloated transactions and OP RETURN data?
Short Answer: This is a very real possibility that we are going to mitigate up front by increasing the average transaction fee (still a very small amount). And we open the amount of this increase up to debate. The best way to explain this is if someone wanted to fill up an 8MB with thousands of small transactions in theory they could. But if it cost them exponentially more to carry this out the attack vector is mitigated. As for OP RETURN data, we will be keep this as is. Future security applications that embed data in OP RETURN should make use of distributed hash tables, not storing data directly on the chain.
Question #5: Transactions are becoming larger than 250 kb, wont this make the issue worse?
Short Answer: Yes some transactions are indeed larger than 250 KB and the average transaction size in the future will undoubtedly be larger but for now its a solid number to work from for averages.
Question #6: With such fast block times and sizes wont transactions be "lost"?
Short Answer: With multi-algorithm mining we are actually in a unique position compared to other coins who have tried with very fast block times. Even though the DigiByte network as a whole will see 15 second block times, each individual algo and mining pool actually is generating 1 minute 15 second blocks. Meaning each block has over a minute to pick up transactions. Some great data on network propagation can be found here: Microsoft Research:
http://www.tik.ee.ethz.ch/file/49318d3f56c1d525aabf7fda78b23fc0/P2P2013_041.pdfQuestion #7: Are you testing massive swings in difficulty?
Short Answer: Absolutely. But after DigiShield and Multi shield this is much less of an issue than the Bitcoin test data that was shown for BTC that showed this to be a problem with larger blocks. Bitcoin still retargets every 2 weeks.
Question #8: Where is the test data for DigiSpeed?
Short Answer: Before releasing the hard fork wallets we will release data on the following points:
*Transaction Times
*Block Times
*Algo Distribution
*Transaction Sizes
*Other data we find useful
Question #9: Why not release DigiSpeed by using a timestamp in the future or by network consensus instead of making network changes occur at a specific block?
Short Answer: By using a specific block everyone knows exactly when the changes to the network will occur. This dramatically mitigates several attack vectors. It also gives people the choice to opt out ahead of time if they so chose. It is also important to note this will be DigiBytes fourth hard fork. The previous three were carried out successfully in the same manner.
Question #10: With block-speed doubling wont that mean twice as many DigiBytes will be coming in to circulation each day?
Short Answer: No. At the hardfork point the reward will be cut in half so the same amount of daily DigiBytes will be mined.
In closing it is important to note we are not going to rush anything here until we are 100% happy with our test data and have addressed all of the communities questions. We would also encourage everyone to read this blog post by the Bitcoin developer Mike Hearn:
https://medium.com/@octskyward/on-block-sizes-e047bc9f830Let us know other technical questions you have and as they come in we will post replies below here:
I'm bumping this because the BCT direct link to the original message does not work and I want to share it with others.