[bitcoin-dev] Fraud proofs for block size/weight

[Luke Dashjr via bitcoin-dev]

Despite the generalised case of fraud proofs being likely impossible, there
have recently been regular active proposals of miners attacking with simply
oversized blocks in an attempt to force a hardfork. This specific attack can
be proven, and reliably so, since the proof cannot be broken without also
breaking their attempted hardfork at the same time.

While ideally all users ought to use their own full node for validation (even
when using a light client for their wallet), many bitcoin holders still do
not. As such, they are likely to need protection from these attacks, to ensure
they remain on the Bitcoin blockchain.

I've written up a draft BIP for fraud proofs and how light clients can detect
blockchains that are simply invalid due to excess size and/or weight:

https://github.com/luke-jr/bips/blob/bip-sizefp/bip-sizefp.mediawiki

I believe this draft is probably ready for implementation already, but if
anyone has any idea on how it might first be improved, please feel free to
make suggestions.

Luke

[Bram Cohen via bitcoin-dev]

Some questions:

Does this require information to be added to blocks, or can it work today
on the existing format?

Does this count number of transactions or their total length? The block
limit is in bytes rather than number of transactions, but transaction
number can be a reasonable proxy if you allow for some false negatives but
want a basic sanity check.

Does this allow for proofs of length in the positive direction,
demonstrating that a block is good, or does it only serve to show that
blocks are bad? Ideally we'd like an extension to SPV protocol so light
clients require proofs of blocks not being too big, given the credible
threat of there being an extremely large-scale attack on the network of
that form.


On Wed, Mar 22, 2017 at 1:47 AM, Luke Dashjr via bitcoin-dev <

Post by Luke Dashjr via bitcoin-dev
Despite the generalised case of fraud proofs being likely impossible, there
have recently been regular active proposals of miners attacking with simply
oversized blocks in an attempt to force a hardfork. This specific attack can
be proven, and reliably so, since the proof cannot be broken without also
breaking their attempted hardfork at the same time.
While ideally all users ought to use their own full node for validation (even
when using a light client for their wallet), many bitcoin holders still do
not. As such, they are likely to need protection from these attacks, to ensure
they remain on the Bitcoin blockchain.
I've written up a draft BIP for fraud proofs and how light clients can detect
https://github.com/luke-jr/bips/blob/bip-sizefp/bip-sizefp.mediawiki
I believe this draft is probably ready for implementation already, but if
anyone has any idea on how it might first be improved, please feel free to
make suggestions.
Luke
_______________________________________________
bitcoin-dev mailing list
https://lists.linuxfoundation.org/mailman/listinfo/bitcoin-dev

[Matt Corallo via bitcoin-dev]

It works today and can be used to prove exact size: the key observation is that all you need to show the length and hash of a transaction is the final SHA256 midstate and chunk (max 64 bytes). It also uses the observation that a valid transaction must be at least 60 bytes long for compression (though much of that compression possibility goes away if you're proving something other than "too large").

Post by Bram Cohen via bitcoin-dev
Does this require information to be added to blocks, or can it work today
on the existing format?
Does this count number of transactions or their total length? The block
limit is in bytes rather than number of transactions, but transaction
number can be a reasonable proxy if you allow for some false negatives but
want a basic sanity check.
Does this allow for proofs of length in the positive direction,
demonstrating that a block is good, or does it only serve to show that
blocks are bad? Ideally we'd like an extension to SPV protocol so light
clients require proofs of blocks not being too big, given the credible
threat of there being an extremely large-scale attack on the network of
that form.
On Wed, Mar 22, 2017 at 1:47 AM, Luke Dashjr via bitcoin-dev <

Post by Luke Dashjr via bitcoin-dev
Despite the generalised case of fraud proofs being likely impossible,

there

Post by Luke Dashjr via bitcoin-dev
have recently been regular active proposals of miners attacking with

simply

Post by Luke Dashjr via bitcoin-dev
oversized blocks in an attempt to force a hardfork. This specific

attack

Post by Luke Dashjr via bitcoin-dev
can
be proven, and reliably so, since the proof cannot be broken without

also

Post by Luke Dashjr via bitcoin-dev
breaking their attempted hardfork at the same time.
While ideally all users ought to use their own full node for

validation

Post by Luke Dashjr via bitcoin-dev
(even
when using a light client for their wallet), many bitcoin holders

still do

Post by Luke Dashjr via bitcoin-dev
not. As such, they are likely to need protection from these attacks,

to

Post by Luke Dashjr via bitcoin-dev
ensure
they remain on the Bitcoin blockchain.
I've written up a draft BIP for fraud proofs and how light clients

can

Post by Luke Dashjr via bitcoin-dev
detect

https://github.com/luke-jr/bips/blob/bip-sizefp/bip-sizefp.mediawiki

Post by Luke Dashjr via bitcoin-dev
I believe this draft is probably ready for implementation already,

but if

Post by Luke Dashjr via bitcoin-dev
anyone has any idea on how it might first be improved, please feel

free to

Post by Luke Dashjr via bitcoin-dev
make suggestions.
Luke
_______________________________________________
bitcoin-dev mailing list
https://lists.linuxfoundation.org/mailman/listinfo/bitcoin-dev

[Jorge Timón via bitcoin-dev]

Great stuff, although the ordering of the sections seems a little bit confusing.

I think it would be clearer to put the "Creation of proofs" section
before "Proof verification", maybe even before "Proof format" if a
high level defintion of "full tx size proof" is provided before.

Also, in "For the full-size proof, each transaction should be assumed
to be at a minimum the stripped-size rather than the fixed 60 bytes."
it seems you are referring to a "full-size block proof" as opposed to
a "full size tx proof", perhaps a better term could be "full-weight
block proof" if what you are referring to is the proof of the weight
instead of only the pre-segwit size.

Perhaps some short definitions for "stripped-size proof", "full tx
size proof", "full-size proof" and maybe also "size component" at the
beginning would be enough.

In "Network protocol", "It should not recheck blocks known to be
valid, " does "known to be valid" include the blocks that the peer
told us where valid (with their hash and 0 in the enumerated varint)?
Those could be invalid too if the peer was lying, no?
Do you mean "It should not recheck blocks known to be invalid,"?

Why do you need to have at least one full tx size?

In Rationale you have:
"
Why must a full tx size proof be included?

This is necessary to establish that the claimed block transaction
count is correct.
"

Why do you need to establish that? If you can establish that the
number of transactions is at least N and that N * 60 bytes is greater
than the size/weight limit, isn't it that enough?


On Wed, Mar 22, 2017 at 10:51 PM, Matt Corallo via bitcoin-dev

Post by Matt Corallo via bitcoin-dev
It works today and can be used to prove exact size: the key observation is
that all you need to show the length and hash of a transaction is the final
SHA256 midstate and chunk (max 64 bytes). It also uses the observation that
a valid transaction must be at least 60 bytes long for compression (though
much of that compression possibility goes away if you're proving something
other than "too large").
On March 22, 2017 1:49:08 PM PDT, Bram Cohen via bitcoin-dev

Post by Bram Cohen via bitcoin-dev
Does this require information to be added to blocks, or can it work today
on the existing format?
Does this count number of transactions or their total length? The block
limit is in bytes rather than number of transactions, but transaction number
can be a reasonable proxy if you allow for some false negatives but want a
basic sanity check.
Does this allow for proofs of length in the positive direction,
demonstrating that a block is good, or does it only serve to show that
blocks are bad? Ideally we'd like an extension to SPV protocol so light
clients require proofs of blocks not being too big, given the credible
threat of there being an extremely large-scale attack on the network of that
form.
On Wed, Mar 22, 2017 at 1:47 AM, Luke Dashjr via bitcoin-dev

Post by Luke Dashjr via bitcoin-dev
Despite the generalised case of fraud proofs being likely impossible, there
have recently been regular active proposals of miners attacking with simply
oversized blocks in an attempt to force a hardfork. This specific attack can
be proven, and reliably so, since the proof cannot be broken without also
breaking their attempted hardfork at the same time.
While ideally all users ought to use their own full node for validation (even
when using a light client for their wallet), many bitcoin holders still do
not. As such, they are likely to need protection from these attacks, to ensure
they remain on the Bitcoin blockchain.
I've written up a draft BIP for fraud proofs and how light clients can detect
https://github.com/luke-jr/bips/blob/bip-sizefp/bip-sizefp.mediawiki
I believe this draft is probably ready for implementation already, but if
anyone has any idea on how it might first be improved, please feel free to
make suggestions.
Luke
_______________________________________________
bitcoin-dev mailing list
https://lists.linuxfoundation.org/mailman/listinfo/bitcoin-dev

_______________________________________________
bitcoin-dev mailing list
https://lists.linuxfoundation.org/mailman/listinfo/bitcoin-dev

[Luke Dashjr via bitcoin-dev]

Post by Jorge Timón via bitcoin-dev
I think it would be clearer to put the "Creation of proofs" section
before "Proof verification", maybe even before "Proof format" if a
high level defintion of "full tx size proof" is provided before.

Creation of proofs isn't part of the spec itself.
I've moved it out of the Specification section (but it's still below it).

Post by Jorge Timón via bitcoin-dev
Also, in "For the full-size proof, each transaction should be assumed
to be at a minimum the stripped-size rather than the fixed 60 bytes."
it seems you are referring to a "full-size block proof" as opposed to
a "full size tx proof", perhaps a better term could be "full-weight
block proof" if what you are referring to is the proof of the weight
instead of only the pre-segwit size.
Perhaps some short definitions for "stripped-size proof", "full tx
size proof", "full-size proof" and maybe also "size component" at the
beginning would be enough.

Added a definitions section above.

Post by Jorge Timón via bitcoin-dev
In "Network protocol", "It should not recheck blocks known to be
valid, " does "known to be valid" include the blocks that the peer
told us where valid (with their hash and 0 in the enumerated varint)?
Those could be invalid too if the peer was lying, no?
Do you mean "It should not recheck blocks known to be invalid,"?

Right, fixed.

Post by Jorge Timón via bitcoin-dev
Why do you need to have at least one full tx size?
"
Why must a full tx size proof be included?
This is necessary to establish that the claimed block transaction
count is correct.
"
Why do you need to establish that? If you can establish that the
number of transactions is at least N and that N * 60 bytes is greater
than the size/weight limit, isn't it that enough?

The only way to establish the number of transactions at all, is to show that a
leaf is a parsable transaction. Which this doesn't actually show, so it's
broken. :( Need to think on this. Any ideas? :/

Luke

[Chris Pacia via bitcoin-dev]

On Mar 25, 2017 12:17 AM, "Luke Dashjr via bitcoin-dev" <bitcoin-***@lists.
linuxfoundation.org> wrote:

Any ideas? :/


Can't the size be aggregated up the tree such that each midstate hash is
the hash of branches below plus the agreegate of the sizes below.

This would make the root hash(left + right + size/weight) and the proof
would just be the preimage.

[Matt Corallo via bitcoin-dev]

I dont think thats true? Sure, you have to assume the block is valid
aside from a too-large size, but it seems sane.

You don't strictly need to show that a leaf is a parseable transaction,
as long as you can assume that the block is valid and that you cannot
forge a SHA256 midstate which, when combined with data with a given
length tag, would result in a hash of a given value (this is a pretty
strong assumption, IMO, IIRC this was not a studied nor a claimed
feature of SHA256).

The only issue is that, since parts of the merkle tree are repeated, you
need to be sure that the counting for minimum number of transactions is
accurate, though I did not review your proposal text to check that.

On 03/25/17 05:16, Luke Dashjr wrote:
- snip -

Post by Luke Dashjr via bitcoin-dev
The only way to establish the number of transactions at all, is to show that a
leaf is a parsable transaction. Which this doesn't actually show, so it's
broken. :( Need to think on this. Any ideas? :/
Luke