On Wed, Apr 27, 2011 at 11:18:23PM +0800, ?????? wrote: > hi, Sebastian > > thx for your time! > comments inline:D > > On Wed, Apr 27, 2011 at 4:22 PM, Sebastian Kiesel <[email protected]>wrote: > > > Hi, > > > > On Wed, Apr 27, 2011 at 09:15:11AM +0800, ?????? wrote: > > > hi, everyone: > > > > > > have the ALTO considered the use cases of wireless, which means how to > > make > > > the selection of the peers in the wireless environment? > > > > the ALTO protocols and procedures feature mechanisms that can be > > leveraged in a mobile / wireless scenario. IMO there are two issues: > > what is IMO stands for? > > 1. When a client is moving between different sections of an access > > network it may have to find a new ALTO server which is able > > to give guidance that is useful at the new client position. > > draft-kiesel-alto-3pdisc-05, sec. 1.3 states: > > > > o A change of the IP address at an interface invalidates the result > > of the ALTO server discovery procedure. For instance, if the IP > > address assigned to a mobile host changes due to host mobility, it > > is required to run the ALTO server discovery procedure for the new > > IP address without relying on earlier gained information. > what does the IP address refer to in "it is required to run the ALTO > server discovery procedure for the new IP address without relying on earlier > gained information"? the new IP address of the roaming user which may > allocated by GGSN or wifi AP? however, that allocation process has nothing > to do with the ALTO server discovery. or it is not correct of my > understanding ?
if the client (peer) notices that it has received a new ip address from dhcp,ggsn,etc. it has to re-run the ALTO server discovery, because getting a new address might be the result of a (significant) movement. > > 2. The actual strategy (algorithm) for selecting peers is out of scope > > of the standardization process, as there are many different scenarios > > with different constraints and optimization goals (e.g., in a given > > mobile access network, is the backhaul capacity an issue or not?). > > can i understand "how to make the selection is the issues decided by P2P > > application"? > > The ALTO client protocol together with a procedure for defining new > > Host Characteristics Attributes (cost types) is able to convey the > > input for a variety of different selection algorithms. ` > do you mean the selection algorithms can be self-defined by client itself? yes. the ALTO protocol specification only specifies the protocol for requesting and receiving information from the ALTO server. This data can be used for influencing the peer selection. But the actual algorithm to do the selection is not being standardized, as this is highly application specific. regards sebastian _______________________________________________ alto mailing list [email protected] https://www.ietf.org/mailman/listinfo/alto
