In a scenario in which some of the parties have more data / compute power / trust reputation (or other meaningful properties) than other parties - are there known methods to do secure multiparty computation with load balancing that takes into account the above-mentioned (or other) parameters?

  • $\begingroup$ Great question. I am almost sure I've seen this sort of thing, but I don't remember which papers it was in. I'll keep looking, but I believe this sort of thing exists. $\endgroup$ – mikeazo Nov 13 '19 at 13:48
  • $\begingroup$ @mikeazo, I'm not sure that paper like this one from Ueli Mauer and Martin Hirt, et al. presents a model in this sense... "Graceful Degradation in Multi-Party Computation". I remember they also have some other papers that extends the adversarial model where we have only: semi-honest or fully malicious... Just my 1 cent of guess. $\endgroup$ – Crypto Learner Nov 16 '19 at 18:16
  • 1
    $\begingroup$ This depends very much on the property you want to "load balance" on. For example, if certain parties have more trust, maybe look at MPC for more expressive access structures than a simple threshold. For more compute power, maybe something like outsourced MPC is relevant. For more data I am not sure, but some protocols are naturally asymmetric in load (e.g. garbled circuits), so may this can be leveraged. $\endgroup$ – Guut Boy Feb 26 at 12:14
  • 1
    $\begingroup$ At least in Garbled Circuit 2-Party computation there's typically one party ("the evaluator") with less computational load than the other, typically on the order of 2x-3x. Though this is no easily tunable parameter. $\endgroup$ – SEJPM Feb 26 at 13:42

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.