Hello! It will use Rendezvous hashing of keys: https://en.wikipedia.org/wiki/Rendezvous_hashing
Regards, -- Ilya Kasnacheev пт, 9 нояб. 2018 г. в 16:24, Alejandro Santos <[email protected]>: > Dear Denis, > > How does Ignite provides load balancing between nodes? > > Is it hash-based? > Is it dynamically allocated? > Is it something else? > > Thanks, > > On Thu, Nov 8, 2018 at 1:14 AM Denis Magda <[email protected]> wrote: > > > > Hi, > > > > In general, the load is balanced because the data is distributed evenly > across a cluster of machines. For instance, if you utilize key-value calls > then each request goes to a specific node. If you're on SQL then a query > might be broadcasted or sent to a specific node as well. > > > > Overall, yes, Ignite is the right solution if you need to scale and > accelerate performance. > > > > -- > > Denis > > > > On Tue, Nov 6, 2018 at 6:55 AM Alejandro Santos <[email protected]> > wrote: > >> > >> Hi all, > >> > >> I've been reading the Ignite documentation and have some technical > >> questions. I need to evaluate massive storage systems for some > >> specific application and I would like to understand how ignite works. > >> > >> My application needs a buffering space that write arbitrary values, > >> but then reads on average half of the values at most once. This is a > >> random process, and we can't really predict which keys will be read. > >> > >> Is ignite the right tool for this application? Do you need more > information? > >> > >> Thank you, > >> > >> -- > >> Alejandro Santos > > > > -- > Alejandro Santos >
