Dear authors, I agree with Wendy's opinion about "Let applications be aware of the different roles" in the bottom of her review. I think it could be concluded into:
1. ALTO should support several cost metrics 2. client should send different request, which determines the cost metric, based on its role I think ALTO is mainly about low-level network information and should not have such functionality. Best Regards, Shenshen Chen 2016-07-12 17:24 GMT+08:00 Stefan HOMMES <[email protected]>: > Hello Chen, > > > > thanks a lot for providing us a feedback to our draft. Here my comments to > your mail: > > > > 1. “Be more general-purpose for block chain”: That’s a good point. > Indeed, the draft should not focus only on the Bitocin network but on > blockchain in general, which was the reason why I have removed the Bitcoin > part from the introduction and described it in a more general way. We will > have a look on other chains as well. > > > > 2. “Consider about security”: A good point as well. I am not sure > how we should tackle this issue for the moment. > > > > 3. “Let applications aware of the different roles”: In my opinion, > there are two ways on how ALTO could answer to a request from a peer: (1) > Either ALTO sends an information that is specific to the role of the peer > that send the request, or (2) it proposes an answer that is independent of > the role, meaning the answer can be used by all types of peers and the peer > decides further based on his role. The second way reduces also the > information that is provided to ALTO, which might be an advantage with > regards to privacy. But I think that ALTO should have some information > about the roles, in order to propose different routes by applying different > kind of metrics. Of course the roles should be as general as possible, and > not be linked to the Bitcoin networks since it is only one example of such > a network. Is it foreseen to have such a functionality in ALTO? > > > > Kind Regards, > > Stefan > > > > > > *From:* Shenshen Chen [mailto:[email protected]] > *Sent:* 11 July 2016 14:52 > *To:* Stefan HOMMES <[email protected]> > *Cc:* [email protected] > *Subject:* Re: [alto] Draft: Alto for the blockchain > > > > Dear authors, > > > > I read your draft and get interested in utilizing ALTO for block chain. I > have considered about it over the past few days. Here are some comments: > > > > 1. Be more general-purpose for block chain > > > > I noticed that there’re some parts of introduction about bitcoin have been > removed from the first version and believed this draft is not specified for > bitcoin. Also mentioned in title, this draft is about block chain. But it > only mentioned bitcoin as use case of block chain. > > > > In bitcoin wiki, it said “Block chains were invented specifically for the > Bitcoin project but they can be applied anywhere a distributed consensus > needs to be established in the presence of malicious or untrustworthy > actors.” > > > > Despite it may cost too much to implement block chain for non-financial > cases, mentioning some other use cases (e.g. alternative chain) could make > the draft looks more general-purpose for block chain. > > > > 2. Consider about security > > > > Compare to traditional data base, I thought the block chain was invented > to solve the security problem. So, it is necessary to consider about > security which is much more important than efficiency. For example, ALTO > server may accept some suboptimal results to improve the randomness and > avoid the prediction mentioned in section 7. > > > > 3. Let applications aware of the different roles > > > > In section 6, it said “This requires that the ALTO server is aware of the > different roles”. But roles like wallet, miner and relay nodes are > specified for bitcoin, it may have different roles in other use cases of > block chain. Including such information in ALTO protocol is not > general-purpose for block chain. > > > > By the other hand, ALTO is invented to provide low-level network > information which application can’t get before. Since the information about > different roles is not low-level, it could be more appropriate for ALTO to > let applications maintain the map relationship between each node and its > role. > > > > Hope these helps. :) > > > > Best Regards, > > Shenshen Chen > > > > > > 2016-07-07 23:36 GMT+08:00 Stefan HOMMES <[email protected]>: > > Dear ALTO group, > > > > I am a research associate from the University of Luxembourg, and we have > submitted a draft that is using ALTO for the blockchain: > > https://datatracker.ietf.org/doc/draft-hommes-alto-blockchain/ > > > > We are very curious and interested to receive some feedback about this > draft. Please feel free to send us your comments. We highly appreciate your > opinion and looking forward to the IETF meeting in Berlin. > > > > Kind Regards, > > Stefan > > > > ------------------------------------------------------------------------- > > Dr. Stefan Hommes > Research Associate, SEDAN team, Room C003 > > Mail: [email protected] > > Phone: (+352) 46 66 44 5834 > > University of Luxembourg > > Interdisciplinary Centre on Security Reliability and Trust (SnT) > > 4, Rue Alphonse Weicker, L2721 Luxembourg > > ------------------------------------------------------------------------- > > > > > _______________________________________________ > alto mailing list > [email protected] > https://www.ietf.org/mailman/listinfo/alto > > >
_______________________________________________ alto mailing list [email protected] https://www.ietf.org/mailman/listinfo/alto
