Reputation: 61
I am working on building a system which involves reading up to 100 keys in batch and have to adhere to strict latency constraints like p50 <= 0.2ms, p90 <= 0.8ms
The total number of keys for me is very small and will remain the same for the foreseeable future due to which the best way for me is to create n standalone redis, each with the same data and for reads, simply load balance on client and perform mget with all the keys to fetch.
I have also tried sharding based solutions (using aerospike instead) but the latency increases.
One issue with this solution is single point of failure. Can someone please provide any cluster configuration that redis natively provides that can be used to solve this problem.
Upvotes: 0
Views: 61