summaryrefslogtreecommitdiff
path: root/bip-0105.mediawiki
blob: 7aaef0445f80ad49743ff085c30f92f740988663 (plain)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
<pre>
  BIP: 105
  Title: Consensus based block size retargeting algorithm
  Author: BtcDrak <btcdrak@gmail.com>
  Status: Draft
  Type: Standards Track
  Created: 2015-08-21
</pre>

==Abstract==

A method of altering the maximum allowed block size of the Bitcoin protocol 
using a consensus based approach.

==Motivation==

There is a belief that Bitcoin cannot easily respond to raising the 
blocksize limit if popularity was to suddenly increase due to a mass adoption 
curve, because co-ordinating a hard fork takes considerable time, and being 
unable to respond in a timely manner would irreparably harm the credibility of 
bitcoin.

Additionally, predetermined block size increases are problematic because they
attempt to predict the future, and if too large could have unintended 
consequences like damaging the possibility for a fee market to develop 
as block subsidy decreases substantially over the next 9 years; introducing 
or exacerbating mining attack vectors; or somehow affect the network in unknown
or unpredicted ways. Since fixed changes are hard to deploy, the damage could be
extensive.

Dynamic block size adjustments also suffer from the potential to be gamed by the
larger hash power.

Free voting as suggested by BIP100 allows miners to sell their votes out of band
at no risk, and enable the sponsor the ability to manipulate the blocksize. 
It also provides a cost free method or the larger pools to vote in ways to
manipulate the blocksize such to disadvantage or attack smaller pools.


==Rationale==

By introducing a cost to increase the block size ensures the mining community 
will collude to increase it only when there is a clear necessity, and reduce it
when it is unnecessary. Larger miners cannot force their wishes so easily
because not only will they have to pay extra a difficulty target, then can be
downvoted at no cost by the objecting hash power.

Using difficulty as a penalty is better than a fixed cost in bitcoins because it
is less predictable.

In order to prevent miners having complete control over blocksize, an upper
limit is required at protocol level. This feature ensures full nodes retain
control over consensus, remembering full nodes are the mechanism to keep miners
honest.


==Specification==

The initial block size limit shall be 1MB.

Each time a miner creates a block, they may vote to increase or decrease the
blocksize by a maximum of 10% of the current block size limit. These votes will 
be used to recalculate the new block size limit every 2016 blocks.

Votes are cast using the block's coinbase field.

The first 4 bytes of the coinbase field shall be repurposed for voting as an
unsigned long integer which will be the block size in bytes.

If a miner votes for an increase, the block hash must meet a difficulty target
which is proportionally larger than the standard difficulty target based on the
percentage increase they voted for.

Votes proposing decreasing the block size limit do not need to meet a higher 
difficulty target.

Miners can vote for no change by voting for the current block size.

For blocks to be valid the blockhash must meet the required difficulty target
for the vote otherwise the block is invalid and will be rejected.

Every 2016 blocks, the block size limit will be recalculated by the median of
all votes in the last 2016 blocks. This will redefine the block size limit for
the next 2016 blocks.

Blocks that are larger than the calculated base block size limit are invalid and
will be rejected.

The base block size limit may not reduce below 1MB or increase above 8MB.


==Acknowledgements==

This proposal is based on ideas and concepts derived from the writings of
Meni Rosenfeld and Gregory Maxwell.


==Copyright==

This work is placed in the public domain.