Re: Atomic Compare and Swap

2010-06-23 Thread Rishi Bhardwaj
Hmnn... vector clocks seem very interesting, I am not very familiar with the work being done in 0.7 for the vector clock support but it sounds like a great addition. I was thinking if we can just use vector clocks to implement CAS in the following way: Let's assume that a client will only do on

Re: Atomic Compare and Swap

2010-06-23 Thread Mike Malone
On Wed, Jun 23, 2010 at 4:51 AM, aaron morton wrote: > I've been playing with something like CAS, it's not the same but it may be > of interest. > > I write some data into Cassandra with quorum or better consistency, that > allows me to assert what it should look like when read back. If the > asse

Re: Atomic Compare and Swap

2010-06-23 Thread aaron morton
I've been playing with something like CAS, it's not the same but it may be of interest. I write some data into Cassandra with quorum or better consistency, that allows me to assert what it should look like when read back. If the assertion holds I can then go ahead. For example, in a CF wi