< All blogs
minute read

How your cloud environment sizing could be hurting FairCom DB's throughput

Blog isometric illustration

Here's a quick tip on maximizing your performance

Did you know? You can host FairCom DB servers in any cloud—public or private. In fact, FairCom customers are already running FairCom server processes through all major cloud vendors.

When working within the cloud, you may want to put your FairCom server module in a docker container to simplify cloud management. To help with this, we've added tips and tricks to our improved guide to administering your docker container.

One big way to improve administration is by managing your cloud environment sizing. We've previously seen customers receive slow throughput even when running on reasonably sized machines. It initially proved quite the problem: The throughput would sometimes be responsive, reaching tens of thousands of transactions per second (tps). At other times, that number would dwindle to a few thousand tps.

After a little time and a lot of investigation, we identified the issue: The computer instance used a network card with a reported bandwidth of "Up to X Gbps" (giga-bits per second). The client needed a different cloud environment sizing. Securing a larger instance with a dedicated network card speed solved the throughput problem.

The problem with the "Up to X Gbps" card is that the computer network could often saturate the card, limiting the FairCom servers' ability to communicate with client processes. When determining cloud environment sizing for FairCom servers, optimize the disk I/O subsystem first. If the client and server run on different machines, optimize the network card (NIC). Then, make sure there's plenty of memory and an adequate CPU

Of course, there are exceptions to every rule, but these tips should get you off to a good start in most circumstances. Want to take a deeper look at your throughput? Try our scriptable performance monitoring tools or our browser-based tools.

Are you interested in taking your FairCom-based product to the cloud? Give our support team a call. We'd enjoy an opportunity to discuss any questions you might have.

Written by:
Randal Hoff
Last Update:
July 2, 2024
FairCom DB