Skip to content

Recommended Operating Environment

Component Recommendation
Cluster Size Sized according to throughput needs — see Cluster Sizing
Operating System Ubuntu 22.04 LTS (recommended)
Quine Enterprise Java OpenJDK or Oracle 21 (requires at least 17)
Quine Enterprise CPU 8-32 Cores Recommended
Quine Enterprise Memory 16-20 GB RAM
Cassandra Java OpenJDK or Oracle 11 (or 8)
Cassandra CPU 16+ Cores
Cassandra Memory 32+ GB RAM; high memory instances are recommended

EC2 instance‑type guidance

Use fixed‑performance instances such as the m7a family for production clusters. m7a nodes deliver sustained vCPU capacity.

Avoid burstable t‑series (t3, t4g, etc.) in production. These instances rely on CPU‑credit buckets: once the credit balance is depleted, the vCPUs are throttled down to their low baseline (for example, about 30% of a core), causing sudden latency spikes and reduced throughput.

Burstable nodes are fine for ad‑hoc testing or dev workloads that idle most of the time, but Quine Enterprise's stream‑processing benefits from guaranteed compute.

Cluster clock synchronization

Important! System clocks should be synchronized between cluster members.

See Clock Synchronization for more details.

Reference deployment diagram, subtitled throughput drives sizing: two Kafka topics feed two Quine Enterprise members and a hot spare, which store data in two Cassandra nodes. Minimum: 1 or more topics with 1 or more partitions; for resilience, a 3-machine Quine Enterprise cluster (2+1) with 4 CPU cores, 8 GB memory, no disk requirement, and Java 17; for high availability, a 2-machine Cassandra cluster with 8 cores, 8 GB memory, 1 TB disk, and replication above 1. Recommended: any number of topics, 1 or more partitions per Quine consumer, and data partitioned for locality; for performance, n Quine Enterprise machines sized by throughput with 16 cores, 16-24 GB memory, no disk requirement, and Java 21; for speed and recovery, n Cassandra machines sized by throughput with 16 cores, 32 GB or more memory, disk sized by throughput and retention, replication 3, and multiple availability zones.