[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: Encoding for Robust Immutable Storage (ERIS)
|
From: |
pukkamustard |
|
Subject: |
Re: Encoding for Robust Immutable Storage (ERIS) |
|
Date: |
Sun, 26 Jul 2020 19:28:49 +0200 |
|
User-agent: |
mu4e 1.4.10; emacs 26.3 |
Hello Christian,
Thank you for your comments!
For my taste, the block size is much too small. I understand 4k
can make
sense for page tables and SATA, but looking at benchmarks 4k is
still
too small to maximize SATA throughput. I would also worry about
4k for a
request size in any database or network protocol. The overheads
per
request are still too big for modern hardware. You could easily
go to
8k, which could be justified with 9k jumbo frames for Ethernet
and would
at least also utilitze all of the bits in your paths. The 32k
of ECRS
are close to the 64k which are reportedly the optimum for modern
M.2
media. IIRC Torrents even use 256k.
I agree that increasing block size makes sense for improving
performance
in storage and transport.
The overhead from padding may be
large for very small files if you go beyond 4k, but you should
also
think in terms of absolute overhead: even a 3100% overhead
doesn't
change the fact that the absolute overhead is tiny for a 1k
file.
The use-case I have in mind for ERIS is very small pieces of data
(not
even small files). Examples include ActivityStreams objects or
OpenStreetMaps nodes.
Apparently the average size of individual ActivityStreams objects
is
less than 1kB (unfortunately I don't have the data to back this
up).
I agree that the overhead of 3100% for a single 1kB object is
acceptable. But I would argue that an overhead of 3100% for very
many
1kB objects is not. The difference might be a 32 GB database
instead of
a 1 GB database.
Furthermore, you should consider a trick we use in GNUnet-FS,
which is
that we share *directories*, and for small files, we simply
_inline_ the
full file data in the meta data of the file that is stored with
the
directory or search result. So you can basically avoid having to
ever
download tiny files as separate entities, so for files <32k we
have zero
overhead this way.
That makes a lot of sense.
But packing multiple objects into a single transport packet or
grouping
for storage on disk/in database works for small block sizes as
well. The
optimization just happens at a "different layer".
The key value I see in having small block sizes is that tiny
pieces of
data can be individually referenced and used (securely).
I'd be curious to see how much the two pass encoding costs in
practice
-- it might be less expensive than ECRS if you are lucky
(hashing one
big block being cheaper than many small hash operations), or
much more
expensive if you are unlucky (have to actually read the data
twice from
disk). I am not sure that it is worth it merely to reduce the
number of
hashes/keys in the non-data blocks. Would be good to have some
data on
this, for various file sizes and platforms (to judge IO/RAM
caching
effects). As I said, I can't tell for sure if the 2nd pass is
virtually
free or quite expensive -- and that is an important detail.
Especially
with a larger block size, the overhead of an extra key in the
non-data
blocks could be quite acceptable.
I think the cost of the two-pass encoding in ERIS is quite
expensive.
Considering that the hash of the individual blocks also needs to
be
computed (as reference in parent nodes), I think ECRS will always
win
performance wise.
Maybe the answer is not ECRS or ERIS but ECRS and ERIS. ECRS for
large
pieces of data, where it makes more sense to have large block size
and
single-pass encoding. And ERIS for (very many) small pieces of
data
where a 3100% overhead is too much but the performance penalty is
acceptable and size of data is much smaller than memory.
There might be some heuristic that says: If data is larger than
2MB use
ECRS, else use ERIS and you get the verification capability.
If using ECRS, you can add the verification capability by encoding
a
list of all the hash references to the ECRS block with ERIS. The
ERIS
read capability of this list of ECRS block is enough to verify the
integrity of the original ECRS encoded content (without revealing
the
content).
What do you think?
For 3.4 Namespaces, I would urge you to look at the GNU Name
System
(GNS). My plan is to (eventually, when I have way too much time
and
could actually re-do FS...) replace SBLOCKS and KBLOCKS of ECRS
with
basically only GNS.
I have been looking into it. It does seem to be a perfect
application of
GNS.
The crypto is way above my head and using readily available and
already
implemented primitives would make implementation much easier for
me. But
I understand the need for "non-standard" crypto and am following
the
ongoing discussions.
-pukkamustard
On 7/10/20 8:59 AM, pukkamustard wrote:
Hello GNUNet,
I'd like to request feedback, questions and comments on an
encoding of
content very much inspired by ECRS that I have been working on:
Encoding
for Robust Immutable Storage (ERIS)
https://openengiadina.net/papers/eris.html
The motivation is to use the encoding in a social network like
settings
where short messages and interactions are encoded using ERIS
(as RDF
[1]).
There is one major difference to ECRS (and a couple smaller
ones) that I
would like to highlight:
** Verification capability
ERIS adds a verification capability. Holders of the
verification
capability can enumerate all blocks required to decode the
content and
verify integrity of the blocks without being able to decode the
content.
This enables peers to cache the entire content without being
able to
read the content.
The verification capability is enabled by using two keys:
1. A read key to encode the blocks holding content.
2. A verification key (which is deterministically derived from
the read
key) to encode the intermediary nodes of the Merkle tree.
This makes the scheme slightly more complicated than ECRS and
also
requires a two-pass encoding (when using convergent
encryption).
Nevertheless I believe this is a very important feature that
maybe
results in a better privacy/complexity/availability trade-off
as alluded
to in a previous thread
(https://lists.gnu.org/archive/html/gnunet-developers/2020-05/msg00015.html).
** Block size
Block size is chosen to be 4kB. This an optimization towards
small
content (short messages and social interactions).
** URN
Encoded content can be referred to by a URN making it usable
from
existing Web (and RDF) settings. This could be added to ECRS.
** No namespacing / keyword search
There are currently no SBlock or KBlock like features. The idea
is that
these features can be built on-top of the base encoding
(including
SBlock and KBlock).
We have a little JavaScript demo:
https://openengiadina.gitlab.io/js-eris/ . As well as
implementation in
Guile [2].
I'd be very happy for your insight and feedback.
Thanks!
-pukkamustard
[1]
https://openengiadina.net/papers/content-addressable-rdf.html
[2] https://gitlab.com/openengiadina/data-model/