BlockDAG Throughput

Posted on 2018-05-12 | Edited on 2018-06-01 |

To compare BlockDAG throughput with blockchain, I take Bitcoin and Etherum as the references in the following table, part of dag’s data comes from this presentation: https://youtu.be/57DCYtk0lWI at 32’.

type	$k$	$\mathbf{D}_{max}$	$δ$	$λ$	throughput	bandwidth	storage	block size	tx size
Bitcoin	0	15s	4.88%	1 Blk/600s	~3.33 txn/s	~1.667 KB/s	~0.15 GB/Day	1 MB	~0.5 KB
Ethereum	0	15s	86.47%	1 Blk/15s	~15 txn/s	~1.667 KB/s	~0.15 GB/Day	~25 KB	~0.1 KB
dag A	354	15s	1.0‰	10 Blks/s	1k txn/s	0.5 MB/s	~43.2 GB/Day	50 KB	0.5 KB
dag B	1621	15s	1.0‰	50 Blks/s	10k txn/s	5 MB/s	~432 GB/Day	100 KB	0.5 KB
dag C	48	15s	1.0‰	1 Blk/s	1M txn/s	500 MB/s	~43.2 TB/Day	0.5 GB	0.5 KB

BlockDAG Blue Selection Algorithm Demo (Golang)

Posted on 2018-04-30 | Edited on 2018-05-15 |

BlockChain (for example Bitcoin, Etherum, etc.) is just a ‘k=0’ special subtype of BlockDAG, that’s why they suffer from the highly restrictive throughput. DAG is the future!

Blue set selection algorithm is the base of the BlockDAG technology. Thanks Go Playground, to create a new example DAG and run it online to see the DAG blue behaviour, it’s quite easy now. For example, to test a DAG in this figure ‘Fig.4’, just add a piece of codes like this:
Fig.4

BlockDAG Blue Selection Algorithm (Rust)

Posted on 2018-04-25 | Edited on 2018-05-15 |

BlockDAG blue set selection algorithm is the base of the blockDAG technology. The ‘blue set’ means a k-cluster SubDAG, let’s use S denotes it, such that |anticone(B)| ∩ S <= k, for all B ∈ S. We term this selection of a k-cluster a coloring of the DAG, and use the colors ‘blue’ and ‘red’ as a convention for blocks inside and outside the chosen cluster, respectively.

A special subtype of BlockDAG is the ‘0-cluster’ SubDAG, that is, BlockChain! Such as Bitcoin, Etherum, and so on. ‘k=0’ is the exact root cause of BlockChains have the highly restrictive throughput, for example ‘Bitcoin’ throughput is 3-7 transactions per second (tps). BlockDAG is the BlockChain’s future!

How to build a test?

For example, to run the simulation for the example Fig.3:
Fig.3

1	$ cargo test test_fig3 -- --nocapture

To run the algorithm simulation for an example of generating 1,000,000 random blocks, and execute the blue selection in a real-time calculation: