EIP-4844: Shard Blob Transactions #4844

protolambda · 2022-02-26T01:28:08Z

This EIP introduces shard blob transactions. See EIP abstract.

~~TODO: update discussions-to link~~ (Done)

eth-bot · 2022-02-26T01:28:52Z

All tests passed; auto-merging...

(pass) eip-4844.md

classification
`updateEIP`

passed!

protolambda · 2022-02-26T01:42:23Z

Force pushes log: Got number 4844 unexpectedly, 4843 doesn't exist, had to change the name. Updated discussions link. Fixed a remaining TODO (tx.header -> tx.message to follow naming pattern established in CL).

MicahZoltu · 2022-02-26T06:02:20Z

I'm very suspicious of an EIP numbered 4844 with @lightclient as an author and issue/PR number 4843 completely missing from GitHub. As soon as I figure out how you managed to skip 4843 I will decide your fate.

MicahZoltu

External links and switching EIP links to relative are the only blockers.

EIPS/eip-4844.md

protolambda · 2022-02-26T16:27:30Z

@MicahZoltu thank you for your review. I implemented your suggestions and moved the background links to the eth magicians discussion thread.

MicahZoltu

CI failures:

./_site/EIPS/eip-4844.html

internally linking to ./eip-4490.md, which does not exist (line 941)
EIP-4490

linking to internal hash #Networking that does not exist (line 430)
Networking

For the anchor link, maybe try ###Networking? I have seen people use anchor links in the past in EIPs, but I never looked into how it works. The rendering system uses Jekyll, so that would be where I would start if you want to dig deeper.

eip-4844.md: description exceeds max length of 140 characters

MicahZoltu · 2022-02-27T06:30:23Z

EIPS/eip-4844.md

+This is equal to the theoretical maximum size of a block today (30M gas / 16 gas per calldata byte = 1.875M bytes), and so it will not greatly increase worst-case bandwidth.
+Post-merge, block times are expected to be static rather than an unpredictable Poisson distribution, giving a guaranteed period of time for large blocks to propagate.
+
+The _sustained_ load of this EIP is much lower than alternatives (eg. [EIP-4488](./eip-4488.md) and [EIP-4490](./eip-4490.md)),


Any EIPs you link to are considered "dependencies", and thus require those EIPs be as far along in the EIP process as this one. EIP-4490 isn't an EIP at all, and EIP-4488 is a draft I think (and may not reach final before this). I recommend just describing (very briefly) the alternatives and then including additional links out in the discussions-to thread.

As mentioned previously, EIPs should stand on their own aside from their dependencies, and the EIP document isn't an appropriate place to make comparisons against other proposals aside from a high level, "we could do X, but we decided this is superior because of Y".

…ncies, describe instead

protolambda · 2022-02-28T14:30:10Z

@MicahZoltu please have a look again:

9e7ecd9 removes the (pre)draft dependencies and describes the differences instead
9b7adeb makes the tx type dependencies (all final EIPs) explicit requirements

EIPS/eip-4844.md

lightclient · 2022-02-28T15:31:33Z

EIPS/eip-4844.md

+and `access_list` as in [`EIP-2930`](./eip-2930.md).
+
+[`EIP-2718`](./eip-2718.md) is extended with a "wrapper data", the typed transaction can be encoded in two forms, dependent on the context:
+- Network (default): `TransactionType || TransactionNetworkPayload`, or `LegacyTransaction`


I believe that technically, EIP-2718 txs are encoded as rlp(rlp(tx type || payload)). This essentially means they are encoded as 0xbX byte strings over the wire. Not sure how relevant that is to this conversation here.

Haven't we had this debate before? There is no nested RLP encoding, it is just RLP, and that is only with regards to how they are encoded by the gossip layer, which has nothing to do with this I don't think.

lightclient · 2022-02-28T15:37:30Z

EIPS/eip-4844.md

+```python
+def tx_hash(tx: SignedBlobTransaction) -> Bytes32:
+    # The pre-image is prefixed with the transaction-type to avoid hash collisions with other tx hashers and types
+    return keccak256(BLOB_TX_TYPE + ssz.hash_tree_root(tx.message))


Interesting, why this vs. BLOB_TX_TYPE + hash(tx)[1:]?

IIUC ssz.hash_tree_root lets you do fancy proofs of specific items within an object, rather than needing to provide the entire object.

EIPS/eip-4844.md

djrtwo

nice!

EIPS/eip-4844.md

djrtwo · 2022-02-28T16:29:56Z

EIPS/eip-4844.md

+The `priority_fee_per_gas` and `max_basefee_per_gas` fields follow [EIP-1559](./eip-1559.md) semantics,
+and `access_list` as in [`EIP-2930`](./eip-2930.md).
+
+[`EIP-2718`](./eip-2718.md) is extended with a "wrapper data", the typed transaction can be encoded in two forms, dependent on the context:


If this is to be a general purpose extension to 2718, we might want to strip it out into a separate EIP

Before creating a whole new EIP I want to get feedback in ACD first (there may be different ways we split this EIP).

EIPS/eip-4844.md

djrtwo · 2022-02-28T16:43:52Z

EIPS/eip-4844.md

+    return ecrecover(tx_hash(tx), sig.v, sig.r, sig.s)
+```
+
+### Beacon chain validation


Suggest moving CL code changes out into consensus-specs repo and use "events" and other types of mechanisms to just reference the outer CL functionality

Will work on a draft for the CL side this week. More clarification about the consensus specs and EIP relation would be appreciated.

EIPS/eip-4844.md

djrtwo · 2022-02-28T16:52:11Z

EIPS/eip-4844.md

+- Along other consensus-layer upgrade modifications to `SignedBeaconBlock`, the `beacon_block` gossip topic is updated 
+  and the `beacon_blocks_by_range/2/` and `beacon_blocks_by_root/2/` type tables are updated.
+- Blobs sidecars are relayed through a new global GossipSub topic: `blobs_sidecar`
+- Blobs sidecars are synced through ReqResp protocols: 


Important to define how long these are expected to be served (e.g. on the order of a month?) to put the bound on when this new data can be pruned

djrtwo · 2022-02-28T16:53:04Z

EIPS/eip-4844.md

+
+Instead of putting rollup block data in transaction calldata, rollups would expect rollup block submitters
+to put the data into blobs. This guarantees availability (which is what rollups need) but would be much cheaper than calldata.
+Rollups need data to be available once, long enough to ensure honest actors can construct the rollup state, but not forever.


Thus why we can have different p2p/node pruning rules than the rest of block data

Co-authored-by: lightclient <14004106+lightclient@users.noreply.github.com>

Co-authored-by: Danny Ryan <dannyjryan@gmail.com>

…able

EIPS/eip-4844.md

Pandapip1 · 2022-03-20T22:22:56Z

As soon as I figure out how you managed to skip 4843 I will decide your fate.

@MicahZoltu if someone opens an issue or PR and then gets banned from GitHub, the issue then 404's (see, or rather fail to see #4928). Just so you know for the future :)

MicahZoltu · 2022-03-21T07:07:11Z

Why didn't I even receive an email if that was the case? For 4928 I at least got an email, since the person got banned after the email went out (which is immediate).

Pandapip1 · 2022-03-21T10:35:18Z

Okay, I have no idea then.

@lightclient

* Shard Blob Transactions * EIP-4844: implement review suggestions from Micah * EIP-4844: update (pre-)draft EIP comparisons to not introduce dependencies, describe instead * EIP-4844: list required tx type EIP requirements * EIP-4844: formatting review fixes by @lightclient Co-authored-by: lightclient <14004106+lightclient@users.noreply.github.com> * EIP-4844: clarifications by @djrtwo Co-authored-by: Danny Ryan <dannyjryan@gmail.com> * EIP-4844: clarify blobs downloading requirement, define is_data_available * Update EIPS/eip-4844.md Co-authored-by: lightclient <14004106+lightclient@users.noreply.github.com> Co-authored-by: Danny Ryan <dannyjryan@gmail.com> Co-authored-by: Micah Zoltu <micah@zoltu.net>

protolambda force-pushed the eip-shard-blob-txs branch from ae8db88 to 9382c25 Compare February 26, 2022 01:28

eth-bot added the newEIPFile label Feb 26, 2022

protolambda force-pushed the eip-shard-blob-txs branch 2 times, most recently from aeb7d5e to 8a360a8 Compare February 26, 2022 01:40

Shard Blob Transactions

8a360a8

protolambda changed the title ~~Shard Blob Transactions~~ EIP-4844: Shard Blob Transactions Feb 26, 2022

MicahZoltu suggested changes Feb 26, 2022

View reviewed changes

EIP-4844: implement review suggestions from Micah

84242ee

MicahZoltu suggested changes Feb 27, 2022

View reviewed changes

EIP-4844: update (pre-)draft EIP comparisons to not introduce depende…

9e7ecd9

…ncies, describe instead

protolambda force-pushed the eip-shard-blob-txs branch from 5abdec7 to 9b7adeb Compare February 28, 2022 13:42

EIP-4844: list required tx type EIP requirements

9b7adeb

lightclient reviewed Feb 28, 2022

View reviewed changes

djrtwo reviewed Feb 28, 2022

View reviewed changes

protolambda and others added 2 commits March 1, 2022 23:09

EIP-4844: formatting review fixes by @lightclient

87e846b

Co-authored-by: lightclient <14004106+lightclient@users.noreply.github.com>

EIP-4844: clarifications by @djrtwo

c9a542a

Co-authored-by: Danny Ryan <dannyjryan@gmail.com>

protolambda mentioned this pull request Mar 1, 2022

Ethereum Core Devs Meeting 133 Agenda ethereum/pm#481

Closed

EIP-4844: clarify blobs downloading requirement, define is_data_avail…

fc4877d

…able

MicahZoltu approved these changes Mar 2, 2022

View reviewed changes

eth-bot added updateEIP and removed newEIPFile labels Mar 2, 2022

eth-bot enabled auto-merge (squash) March 2, 2022 09:52

MicahZoltu reviewed Mar 2, 2022

View reviewed changes

EIPS/eip-4844.md Outdated Show resolved Hide resolved

Update EIPS/eip-4844.md

2e3ba31

eth-bot merged commit 5868c8c into ethereum:master Mar 2, 2022

greenlucid mentioned this pull request May 1, 2022

Proposal to add EIP-5022 to Shanghai ethereum/pm#519

Closed

abcoathup mentioned this pull request Jul 20, 2022

How should we pick EIP numbers? #5294

Closed

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

EIP-4844: Shard Blob Transactions #4844

EIP-4844: Shard Blob Transactions #4844

protolambda commented Feb 26, 2022 •

edited

Loading

eth-bot commented Feb 26, 2022 •

edited

Loading

protolambda commented Feb 26, 2022

MicahZoltu commented Feb 26, 2022

MicahZoltu left a comment

protolambda commented Feb 26, 2022

MicahZoltu left a comment

MicahZoltu Feb 27, 2022

protolambda commented Feb 28, 2022

lightclient Feb 28, 2022

MicahZoltu Mar 1, 2022 •

edited

Loading

lightclient Feb 28, 2022

MicahZoltu Mar 1, 2022

djrtwo left a comment

djrtwo Feb 28, 2022

protolambda Mar 1, 2022

djrtwo Feb 28, 2022

protolambda Mar 1, 2022

djrtwo Feb 28, 2022

djrtwo Feb 28, 2022

Pandapip1 commented Mar 20, 2022 •

edited

Loading

MicahZoltu commented Mar 21, 2022

Pandapip1 commented Mar 21, 2022

EIP-4844: Shard Blob Transactions #4844

EIP-4844: Shard Blob Transactions #4844

Conversation

protolambda commented Feb 26, 2022 • edited Loading

eth-bot commented Feb 26, 2022 • edited Loading

All tests passed; auto-merging...

(pass) eip-4844.md

protolambda commented Feb 26, 2022

MicahZoltu commented Feb 26, 2022

MicahZoltu left a comment

Choose a reason for hiding this comment

protolambda commented Feb 26, 2022

MicahZoltu left a comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

protolambda commented Feb 28, 2022

Choose a reason for hiding this comment

MicahZoltu Mar 1, 2022 • edited Loading

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

djrtwo left a comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Pandapip1 commented Mar 20, 2022 • edited Loading

MicahZoltu commented Mar 21, 2022

Pandapip1 commented Mar 21, 2022

protolambda commented Feb 26, 2022 •

edited

Loading

eth-bot commented Feb 26, 2022 •

edited

Loading

MicahZoltu Mar 1, 2022 •

edited

Loading

Pandapip1 commented Mar 20, 2022 •

edited

Loading