Once the 50BTC per block rewards have halved many times and the transaction fees start to become the primary reward for miners, it sound like there is incentive for miners to not pass on transactions as they would eventually get the transaction fee when they next solve a block. I just read about this in Slashdot -
Refers to another article, being: http://coderrr.wordpress.com/2011/11/13/simplified-summary-of-microsoft-researchs-bitcoin-paper-on-incentivizing-transaction-propagation/ and a research paper produced by Microsoft (who would've thought they're interested in Bitcoins!?): http://research.microsoft.com/pubs/156072/bitcoin.pdf
An obvious solution to the described problem is to reward nodes for sharing transactions, but this (as I understand it) leads to a "Sybil attack" where the attacker tries to gain an increased share of the rewards by sharing the transaction amongst other nodes controlled by the attacker.
The rest of the research paper seems to have a proposed solution, but it's beyond my understanding!
Is this really a problem with the existing Bitcoin protocol? If yes, does the research paper have a genuine workable solution? Is it possible to describe that solution in laymen's terms?