From: Tamas Blummer <tamas.blummer@gmail.com>
Content-Type: multipart/alternative;
	boundary="Apple-Mail=_67460D2A-3269-4541-9006-18E24DCBEF89"
Mime-Version: 1.0 (Mac OS X Mail 12.4 \(3445.104.11\))
Date: Sat, 21 Sep 2019 23:16:25 +0200
References: <236E3AB2-4035-4F51-84EE-6F7F57298777@bitaps.com>
To: "admin@bitaps.com" <admin@bitaps.com>,
	Bitcoin Protocol Discussion <bitcoin-dev@lists.linuxfoundation.org>
In-Reply-To: <236E3AB2-4035-4F51-84EE-6F7F57298777@bitaps.com>
Message-Id: <1ACE761E-A91B-45C7-A0EB-BD66FE3AD7BD@gmail.com>
Subject: Re: [bitcoin-dev] Block Batch Filters for Light Clients
Precedence: list


--Apple-Mail=_67460D2A-3269-4541-9006-18E24DCBEF89
Content-Transfer-Encoding: quoted-printable
Content-Type: text/plain;
	charset=utf-8

Hi Aleksey,

Yes, BIP158 uses the block hash to seed the hash function, which makes =
distinct block filters non-aggregatable=20
for common values. Aggregate fiters on ranges of blocks would have to =
use some other seed and then=20
achive significant savings using the same design.

I think that the most likely use of filters is to decide if a newly =
announced block should be downloaded and=20
not scanning over the entire chain, where aggregate filters would help. =
I also suspect that whole chain=20
scans would be better served with plain sequential reads in map-reduce =
style.

Typical clients do not care of filters for blocks before the birth date =
of their wallet=E2=80=99s keys, so they skip over the=20
majority of history which is a bigger saving than any aggregate filter.

I wish we get a filter committed as commitment would unlock more utility =
than any marginal savings through
more elaborate design.

Tamas Blummer

> On Sep 19, 2019, at 19:20, admin--- via bitcoin-dev =
<bitcoin-dev@lists.linuxfoundation.org> wrote:
>=20
> Hello list,=20
>=20
> Here is a link for a draft of a BIP for  compact probabilistic block =
filters alternative of BIP 158
>=20
> =
https://docs.google.com/document/d/1jH9tEUyb9w2OZd4-kxfGuyNIIZzmgkEb_z0qSx=
v80ik/edit?usp=3Dsharing =
<https://docs.google.com/document/d/1jH9tEUyb9w2OZd4-kxfGuyNIIZzmgkEb_z0qS=
xv80ik/edit?usp=3Dsharing>
>=20
> Summary:
>=20
>  - BIP 158  false positive rate is low, we can achieve lower bandwidth =
with higher false positive rate filter while sync blockchain
>=20
>  - BIP 158 not do not support filter batching by design of used =
parameters for siphash and Golomb coding optimal parameters
>=20
>  - Alternative compression with delta coding and splitting data to 2 =
bit string  sequences. First for data without prefixes, second one for =
information about  bit length written to first sequence.
>    Second sequence have a lot of duplicates,  compressed with 2 round =
of Huffman algorithm. (Effectivity about 98% vs Golomb with optimal =
parameters)
>=20
>  - Block filters batching reduce filter size significantly
>=20
> - Separation of filters by address type allows lite client not to =
download redundant information without compromising privacy.
>=20
> - Lite client filters download strategy: get biggest filter (smallest =
blocks/size rate) for blocks range, in case positive test  -> get medium =
filters to reduce blocks range ->  get block filters for affected range =
-> download affected blocks over TOR=20
>=20
> Implementation (python): =
https://github.com/bitaps-com/pybtc/blob/bugfix/pybtc/functions/filters.py=
#L172 =
<https://github.com/bitaps-com/pybtc/blob/bugfix/pybtc/functions/filters.p=
y#L172>
>=20
> Exactly information from mainnet  about size for separated filters by =
address types and batch size will be added within few days.
>=20
> Thanks for any feedback.
>       Aleksey Karpov
>=20
> _______________________________________________
> bitcoin-dev mailing list
> bitcoin-dev@lists.linuxfoundation.org
> https://lists.linuxfoundation.org/mailman/listinfo/bitcoin-dev


--Apple-Mail=_67460D2A-3269-4541-9006-18E24DCBEF89
Content-Transfer-Encoding: quoted-printable
Content-Type: text/html;
	charset=utf-8

<html><head><meta http-equiv=3D"Content-Type" content=3D"text/html; =
charset=3Dutf-8"></head><body style=3D"word-wrap: break-word; =
-webkit-nbsp-mode: space; line-break: after-white-space;" class=3D"">Hi =
Aleksey,<div class=3D""><br class=3D""></div><div class=3D"">Yes, BIP158 =
uses the block hash to seed the hash function, which makes distinct =
block filters non-aggregatable&nbsp;</div><div class=3D"">for common =
values. Aggregate fiters on ranges of blocks would have to use some =
other seed and then&nbsp;</div><div class=3D"">achive significant =
savings using the same design.</div><div class=3D""><br =
class=3D""></div><div class=3D"">I think that the most likely use of =
filters is to decide if a newly announced block should be downloaded =
and&nbsp;</div><div class=3D"">not scanning over the entire chain, where =
aggregate filters would help. I also suspect that whole =
chain&nbsp;</div><div class=3D"">scans would be better served with plain =
sequential reads in map-reduce style.</div><div class=3D""><br =
class=3D""></div><div class=3D"">Typical clients do not care of filters =
for blocks before the birth date of their wallet=E2=80=99s keys, so they =
skip over the&nbsp;</div><div class=3D"">majority of history which is a =
bigger saving than any aggregate filter.</div><div class=3D""><br =
class=3D""></div><div class=3D"">I wish we get a filter committed as =
commitment would unlock more utility than any marginal savings =
through</div><div class=3D"">more elaborate design.</div><div =
class=3D""><br class=3D""></div><div class=3D"">Tamas Blummer</div><div =
class=3D""><br class=3D""></div><div class=3D""><div><blockquote =
type=3D"cite" class=3D""><div class=3D"">On Sep 19, 2019, at 19:20, =
admin--- via bitcoin-dev &lt;<a =
href=3D"mailto:bitcoin-dev@lists.linuxfoundation.org" =
class=3D"">bitcoin-dev@lists.linuxfoundation.org</a>&gt; wrote:</div><br =
class=3D"Apple-interchange-newline"><div class=3D""><meta =
http-equiv=3D"Content-Type" content=3D"text/html; charset=3Dus-ascii" =
class=3D""><div style=3D"word-wrap: break-word; -webkit-nbsp-mode: =
space; line-break: after-white-space;" class=3D"">Hello list,&nbsp;<div =
class=3D""><br class=3D""></div><span class=3D"">Here is a link for a =
draft of a BIP for &nbsp;compact probabilistic block filters alternative =
of BIP 158</span><div class=3D""><br class=3D""></div><div class=3D""><a =
href=3D"https://docs.google.com/document/d/1jH9tEUyb9w2OZd4-kxfGuyNIIZzmgk=
Eb_z0qSxv80ik/edit?usp=3Dsharing" =
class=3D"">https://docs.google.com/document/d/1jH9tEUyb9w2OZd4-kxfGuyNIIZz=
mgkEb_z0qSxv80ik/edit?usp=3Dsharing</a></div><div class=3D""><br =
class=3D""></div><div class=3D"">Summary:</div><div class=3D""><br =
class=3D""></div><div class=3D"">&nbsp;- BIP 158 &nbsp;false positive =
rate is low, we can achieve lower bandwidth with higher false positive =
rate filter while sync blockchain</div><div class=3D""><br =
class=3D""></div><div class=3D"">&nbsp;- BIP 158 not do not support =
filter batching by design of used parameters for siphash and Golomb =
coding optimal parameters</div><div class=3D""><br class=3D""></div><div =
class=3D"">&nbsp;- Alternative compression with delta coding and =
splitting data to 2 bit string &nbsp;sequences. First for data without =
prefixes, second one for information about &nbsp;bit length written to =
first sequence.</div><div class=3D"">&nbsp; &nbsp;Second sequence have a =
lot of duplicates, &nbsp;compressed with 2 round of Huffman algorithm. =
(Effectivity about 98% vs Golomb with optimal parameters)</div><div =
class=3D""><br class=3D""></div><div class=3D"">&nbsp;- Block filters =
batching reduce filter size significantly</div><div class=3D""><br =
class=3D""></div><div class=3D"">- Separation of filters by address type =
allows lite client not to download redundant information without =
compromising privacy.</div><div class=3D""><br class=3D""></div><div =
class=3D"">- Lite client filters download strategy: get biggest filter =
(smallest blocks/size rate) for blocks range, in case positive test =
&nbsp;-&gt; get medium filters to reduce blocks range -&gt; &nbsp;get =
block filters for affected range -&gt; download affected blocks over =
TOR&nbsp;</div><div class=3D""><br class=3D""></div><div =
class=3D"">Implementation (python):&nbsp;<a =
href=3D"https://github.com/bitaps-com/pybtc/blob/bugfix/pybtc/functions/fi=
lters.py#L172" =
class=3D"">https://github.com/bitaps-com/pybtc/blob/bugfix/pybtc/functions=
/filters.py#L172</a></div><div class=3D""><br class=3D""></div><div =
class=3D"">Exactly information from mainnet &nbsp;about size for =
separated filters by address types and batch size will be added within =
few days.</div><div class=3D""><br class=3D""></div><div class=3D"">Thanks=
 for any feedback.</div><div class=3D"">&nbsp; &nbsp; &nbsp; Aleksey =
Karpov</div><div class=3D""><div class=3D""><span class=3D""><br =
class=3D""></span></div></div></div>______________________________________=
_________<br class=3D"">bitcoin-dev mailing list<br class=3D""><a =
href=3D"mailto:bitcoin-dev@lists.linuxfoundation.org" =
class=3D"">bitcoin-dev@lists.linuxfoundation.org</a><br =
class=3D"">https://lists.linuxfoundation.org/mailman/listinfo/bitcoin-dev<=
br class=3D""></div></blockquote></div><br class=3D""></div></body></html>=

--Apple-Mail=_67460D2A-3269-4541-9006-18E24DCBEF89--