Hi, we are using ppe.optimized-pricing-api.refinitiv.com servers for the realtime streaming and we connect using EMA.
Can you indicate us which is the rate of RIC subscriptions that the server accepts per second?
Can we have information about the real-time servers limits ?
If the application sends 2000 RIC/second will the subscriptions be done in the same time or the subscriptions are done in a laps of time?
Thanks for your help,
Regards Cornelia
Hi @Cornelia.Draculet ,
Thank you for your participation in the forum.
Are any of the replies below satisfactory in resolving your query?
If yes please click the 'Accept' text next to the most appropriate reply. This will guide all community members who have a similar question.
Otherwise please post again offering further insight into your question.
Thanks,
AHS
Hi,
Please be informed that a reply has been verified as correct in answering the question, and marked as such.
Thanks,
AHS
This is not an API, but the RTDS/hardware question, and the answer will depend on the type of infrastructure setup you have.
You will have to try and load test this in your particular setup, and if you start seeing bottlenecks, start horizontally scaling the infrastructure and your application.
You can get a hint of what is achievable from the API Performance document.
Its hard to quantify the response rate from the server, since the subscriptions may be fulfilled by different parts of the infrastructure. In general for subscribing to large number of instruments, you should use a batch request - see this example here. The subscription is opened for all instruments at once, but the response messages will arrive in asynchronous manner. The time it takes to fulfill the response depends on where the instrument is trading, how far (hops) you are from data center and what is the market condition (open, closing run etc).
The RTO endpoint your app can connect to, depends on your machine ID, which dictates the maximum number of subscriptions (watchlist) you can open.
Hi, thanks for you answer.
In our case the QA executes performance tests.
They simulate 2500 users that connects to our service, one connection/per second .
Once connected the user sends a calculation request that needs about 500 RIC-s to be subscribed. It can take more than 2 minutes to have all ric-s receiving an image.
We have accesses for 250k subscriptions, but the question is the limitation when we subscribe - is the server accept 100 ric/second, 1000ric/second?
Thanks again,
regards,
Cornelia
On one of the POC we did last year, we were able to scale the infrastructure and use single instance of testclient with 100 RSSL mounts, and got 250K images/second using 400 byte image size.
