Refactor chain state persistency

[ffakenz]

🪚 Remove previous attribute from ChainStateAt.

🪚 Since ChainStateType is not recursive anymore, the HeadState now keeps a non-empty list of ChainStateType instead.

🪚 This means, now the HeadLogic has to collect the new ChainStateType during 'aggregate'.

🪚 Finally, the LocalChainState now has to maintain a non-empty list instead in order to implement 'rollback'.

🪚 As a consequence, now the persistent state is more lightweight and easier to read and work with.

---

• CHANGELOG updated or not needed
• Documentation updated or not needed
• Haddocks updated or not needed
• No new TODOs introduced or explained herafter

[github-actions]

Transactions Costs

Sizes and execution budgets for Hydra protocol transactions. Note that unlisted parameters are currently using arbitrary values and results are not fully deterministic and comparable to previous runs.

Metadata
Generated at	2023-09-08 10:54:49.744790202 UTC
Max. memory units	14000000
Max. CPU units	10000000000
Max. tx size (kB)	16384

Script summary

Name	Hash	Size (Bytes)
νInitial	3ffaf6b87df35cb01a52eb23032b8f0b1a2a3ad3acf0930abc9c833a	4150
νCommit	e4c32d6dc83b2917aa7805571f30437ad98b6d20d821d34d45943755	2093
νHead	8a43c1c4d5cb60c212e7aa540932f311cb914a1b6104f0f36a2aaaf0	8845
μHead	efd460b736e8155861e909d6507760b24a5c28717591c6e98c26b104*	4187

• The minting policy hash is only usable for comparison. As the script is parameterized, the actual script is unique per Head.

Cost of Init Transaction

Parties	Tx size	% max Mem	% max CPU	Min fee ₳
1	4778	11.82	4.65	0.49
2	4985	14.46	5.67	0.53
3	5190	16.87	6.58	0.56
5	5600	21.43	8.31	0.63
10	6623	33.37	12.85	0.80
37	12162	98.57	37.66	1.75

Cost of Commit Transaction

This is using ada-only outputs for better comparability.

UTxO	Tx size	% max Mem	% max CPU	Min fee ₳
1	599	12.67	4.97	0.31
2	787	16.49	6.69	0.37
3	972	20.28	8.39	0.42
5	1351	28.46	12.01	0.53
10	2283	51.35	21.96	0.82
18	3787	95.26	40.52	1.38

Cost of CollectCom Transaction

Parties	UTxO (bytes)	Tx size	% max Mem	% max CPU	Min fee ₳
1	57	815	24.07	9.57	0.45
2	114	1143	38.52	15.40	0.62
3	170	1456	51.90	21.00	0.79
4	228	1784	73.63	29.76	1.04
5	283	2096	91.05	37.07	1.25

Cost of Close Transaction

Parties	Tx size	% max Mem	% max CPU	Min fee ₳
1	689	19.26	8.84	0.40
2	910	21.03	10.44	0.44
3	948	20.12	9.06	0.42
5	1552	25.35	14.68	0.53
10	2590	34.64	22.75	0.71
50	10848	98.49	81.92	2.01

Cost of Contest Transaction

Parties	Tx size	% max Mem	% max CPU	Min fee ₳
1	722	22.75	10.15	0.44
2	928	24.40	11.62	0.47
3	1164	27.15	13.66	0.52
5	1573	30.70	16.75	0.59
10	2696	40.93	25.32	0.78
43	9234	98.57	75.30	1.90

Cost of Abort Transaction

Some variation because of random mixture of still initial and already committed outputs.

Parties	Tx size	% max Mem	% max CPU	Min fee ₳
1	5023	21.53	9.26	0.61
2	5532	36.71	15.98	0.81
3	5931	55.05	24.07	1.03
4	6346	76.47	33.51	1.29

Cost of FanOut Transaction

Involves spending head output and burning head tokens. Uses ada-only UTxO for better comparability.

Parties	UTxO	UTxO (bytes)	Tx size	% max Mem	% max CPU	Min fee ₳
5	0	0	4808	9.06	3.81	0.46
5	1	57	4836	10.37	4.61	0.48
5	5	284	4986	15.36	7.69	0.55
5	10	570	5163	22.18	11.79	0.64
5	20	1140	5524	35.14	19.72	0.82
5	30	1706	5881	47.91	27.56	0.99
5	40	2277	6246	61.31	35.67	1.18
5	50	2847	6602	74.19	43.56	1.35
5	69	3929	7286	99.69	58.99	1.70

[github-actions]

Test Results

358 tests  ±0   353 ✔️ ±0   23m 34s ⏱️ +56s
121 suites ±0       5 💤 ±0 
    6 files   ±0       0 ❌ ±0 

Results for commit fe14efe. ± Comparison against base commit 2340ad5.

♻️ This comment has been updated with latest results.

[v0d1ch]

🎉

[pgrange]

Before, the chainState was really abstract to the HeadLogic but that came with the overhead of having to store the full chainState history with every chain originated event. I'm not sure how heavy this overhead was but I can see how this PR removes it and that's cool.

So now, we do not store the chainState full history every time but it comes with the HeadlLogic having to know about the fact that there is a notion of chainState history and the headLogic know has to deal with it and, for instance, call this pushNewState function.

That's a shame the HeadLogic has to have this knowledge but I understand that there is a trade-off here and I'm fine with this way of dealing with it 👍

[ch1bo]

I like the direction this is going. The fact that we do not need the full history on every chain event is great.

I do have two major points though, which are worth looking into.

As we are refactoring here, I feel like we should be pushing further and directly to a better situation. Details below 👇