1 Answer
- Newest
- Most votes
- Most comments
0
Hey Renz,
Amazon GameLift's FlexMatch feature lets you fully customize your rules, so it can be as simple as you need. Additionally, they provide an intersection expression that might be what you want: set_intersection
(rule expression reference).
Amazon GameLift also provides Standalone FlexMatch which allows you to use their matchmaking solution in isolation and with a different pricing model from standard GameLift making it useful even for non-game applications (check out Pricing > Amazon GameLift FlexMatch for more info).
Hope that helps!
answered a year ago
Relevant content
- asked 6 years ago
- asked 3 years ago
- AWS OFFICIALUpdated 2 years ago
- AWS OFFICIALUpdated 2 years ago
- AWS OFFICIALUpdated 5 months ago
- How do I troubleshoot issues with Amazon Q in Connect real-time recommendations that don't populate?AWS OFFICIALUpdated a month ago
The problem I have with FlexMatch is that it would be super expensive due to Matchmaking Hours. Given that there has to be an intersection of interests between two users, I expect there would be a longer waiting time. I'm thinking of users waiting in the queue for 5 seconds before they're randomly matched to a stranger. If my understanding is correct, this 5 seconds waiting in the queue is billed as 5 seconds in matchmaking hours, if my hypothesis is correct then FlexMatch would be an expensive option.
Hey Renz, FlexMatch pricing can be a little confusing because it depends heavily on the traffic patterns you expect from your users. The key is that price is determined by how long matches take to complete not how long a user takes to get into a match. A single unmatchable user could be tried for 10-20 successful matches before they're matched as long as more users continue to join. If you won't have a lot of concurrent users, or the number of possible interests is high enough that users are unlikely to match quickly, it will be expensive. However, if you expect a consistent number of users and the interests are likely to overlap, you'll tend to see most matches complete quickly regardless of a player taking 5 seconds to match.
You might be able to get away with using SQS to trigger lambda functions with batches of users. Some articles to look into: https://docs.aws.amazon.com/lambda/latest/dg/with-sqs.html, https://aws.amazon.com/blogs/compute/understanding-how-aws-lambda-scales-when-subscribed-to-amazon-sqs-queues/