Hi, Jing Chen wrote: > (1) Do we really need using HIP in P2PSIP? I mean that if we can > seperate the peerID and its location info(such as IP addr) through some > Name Services(e.g. P2PNS), do we need HIP?
Well, we *can* do P2PSIP without HIP, but HIP would make many things easier. After all, HIP is much more than just a realization of the ID-locator split. We get, e.g., mobility and multihoming that works with registration and also during an active session. Also, HIP provides end-to-end secured IPsec connections (if needed; this can also be disabled to save CPU resources), Denial of Service attack protection, cryptographically secure identities etc. Finally, since NAT traversal is quite essential for P2PSIP and there are ongoing efforts on integrating ICE-like NAT traversal [1] for at least two different open-source HIP implementations [2] [3], using HIP can simplify implementing the whole P2PSIP quite a bit. > (2) What's the essential benifits to introducing HIP in P2PSIP? What's > the essential differences between P2PSIP overlay nodeID and the HITs in HIP? So, basically, the essential benefit is that you can leave many things that are hard to implement for HIP and concentrate on the P2PSIP specific things. Some good further reading is e.g., the HIP BONE draft [4]. Cheers, Ari [1] http://tools.ietf.org/id/draft-ietf-hip-nat-traversal-03.txt [2] http://hip4inter.net/ [3] http://infrahip.hiit.fi/ [4] http://tools.ietf.org/id/draft-camarillo-hip-bone-01.txt _______________________________________________ P2PSIP mailing list [email protected] https://www.ietf.org/mailman/listinfo/p2psip
