Message-ID: <554CC19D.2010306@xylon.de>
Date: Fri, 08 May 2015 16:01:01 +0200
From: Arne Brutschy <abrutschy@xylon.de>
User-Agent: Mozilla/5.0 (X11; Linux x86_64;
	rv:31.0) Gecko/20100101 Thunderbird/31.6.0
MIME-Version: 1.0
To: bitcoin-development@lists.sourceforge.net
Content-Type: text/plain; charset=iso-8859-15; format=flowed
Content-Transfer-Encoding: 7bit
Subject: [Bitcoin-development] Removing transaction data from blocks
Precedence: list

Hello,

At DevCore London, Gavin mentioned the idea that we could get rid of 
sending full blocks. Instead, newly minted blocks would only be 
distributed as block headers plus all hashes of the transactions 
included in the block. The assumption would be that nodes have already 
the majority of these transactions in their mempool.

The advantages are clear: it's more efficient, as we would send 
transactions only once over the network, and it's fast as the resulting 
blocks would be small. Moreover, we would get rid of the blocksize limit 
for a long time.

Unfortunately, I am too ignorant of bitcoin core's internals to judge 
the changes required to make this happen. (I guess we'd require a new 
block format and a way to bulk-request missing transactions.)

However, I'm curious to hear what others with a better grasp of bitcoin 
core's internals have to say about it.

Regards,
Arne