1

I have successfully installed PIO on a single machine and I see that the recommendation engine works fine. However, I would like to know how I can scale up the system to distribute the load across multiple machines, and I can't find anywhere clear instructions how to do it. Thanks for your help.

Mister_L
  • 2,469
  • 6
  • 30
  • 64
  • Are you hosting your machine on an amazon? I used amazon's [elastic load balancing](https://aws.amazon.com/elasticloadbalancing/) to spread requests across several predictionio servers – alex9311 Jul 31 '16 at 14:28
  • @alex9311 I can set multiple instances of PIO behind a load balancer, but I don't want to duplicate the events stored, instead I want to distribute both the storage and computational load among all PIO instances. – Mister_L Aug 01 '16 at 20:14
  • @Mister_L: Events won't be duplicated unless you have multiple `Event Servers`. For horizontal scoring -sever scaling, one should deploy models (using `pio deploy`) on multiple AWS node instances and utilize Load Balancer ([ELB](https://aws.amazon.com/elasticloadbalancing/)) on top of it. – CᴴᴀZ Apr 02 '18 at 12:09

0 Answers0