Ripple Node & Ledger History Questions

[micro4lpha]

We are building a small exchange and I am in charge of running the nodes with the proper configurations in order for our application to get all the proper info from the blockchain using our custom built API.

I have installed a "small" node size on an m5.large at AWS and I'm having problem with the server usage. Even though its a large server it struggles a little and running a 32GB RAM server or 9 TB of storage would increase costs considerably. I was told Ripple runs its own servers, is it possible to connect via URL to these in order to not host our own node?

We might consider creating an API using Ripple servers directly rather than running our own as running a full history ledger is about 9 TB. What other solutions could we apply in this case?

Any help is greatly appreciated!

Edited January 13, 2020 by micro4lpha

[at3n]

See https://xrpl.org/get-started-with-the-rippled-api.html#public-servers. Note that it mentions that they're not intended for business use. There have been several instances over the past years where other exchanges which have relied on the public Ripple servers experienced performance issues because the Ripple servers do sometimes have problems. So it's highly recommend to run your own. 

[Sukrim]

Make sure you have high performance SSDs as storage layer, I'm not sure if you can get that from AWS actually...

As someone running a full history server: Yeah, it is expensive. Why do you need data from 2013 though to operate an exchange in 2020?