> > > 2) Trivial "recovery" algorithms as the server can, after a rollback, > > > get operations back from clients > > > Shifting the responsibility for reliability from a trusted server to > > other clients seems like a questionable solution. I think reliably > > storing the server's data shouldn't be the job of clients. > > This is just silly. No one is shifting any responsibility....
I think an important point has been missed here. It is not only clients that can resend operations to a server which has rolled back. With TP2 a number of servers can be connected to one another and keep each other in sync. If a server in the server cluster rolls back, it can be brought up to date by the other servers to which it is connected in the cluster, not only by "clients". Responsibility can be shared not only with the clients, but also amongst the other servers. I fail to see how the Wave approach can offer this kind of reliability and scalability. With the Wave approach, for a given wavelet (I think this terminology is correct), there is a single machine which is responsible, and it is a single point of failure. Nothing can be done if that single physical (or perhaps virtual) machine crashes, except to roll back. Even if it's no fault of the Wave software, operating systems crash, and such crashes, and consequent rollback must be taken care of. In short, the Wave approach is not aided by the "throw another machine at it" approach, whereas a TP2-based approach is. > What is described above is a classic TP2 puzzle. In a peer-to-peer > system this would represent divergence amongst peers. In wave it just > gives a plain weird result. I see it as not preserving the intention > of the users. Dan, this comes across as misleading to me. A TP2 based system will converge in this case (and all others). Importantly, the ordering of AB (and thus the obvious intention) is maintained. I think a misunderstood comment later in this thread was made based on confusion from this point: On Feb 28, 3:22 am, Tad Glines <[email protected]> wrote: ... > While in one of the 6 cases wave would produce "BA" instead of "AB" all > clients would converge on the same result. So no divergence would occur as > it would in many TP2 systems" Tad, as David asks later, on what basis do you suggest that many TP2 systems would diverge, was it based on Dan's comment above? I hope this has clarified a couple of things. Regards, Jesse -- You received this message because you are subscribed to the Google Groups "Wave Protocol" group. To post to this group, send email to [email protected]. To unsubscribe from this group, send email to [email protected]. For more options, visit this group at http://groups.google.com/group/wave-protocol?hl=en.
