11 articles on streaming — what drives the cost, how it is priced, and where the savings actually are.
Pub/Sub prices throughput at a flat rate per TiB with no shards to provision, which is either a bargain or a trap depending on your fan out factor and retention settings. Here is the full meter list.
Change data capture replaces expensive full table reloads with a stream of row level changes, but it introduces its own meters: connector compute, log retention, stream throughput, and merge cost at the sink.
A streaming service with a 10,000 hour library and a million viewing hours a month costs about 33,680 dollars on AWS, and 89 percent of that is CDN egress. Here is the breakdown, and why bitrate is the only number that really matters.
A managed Kafka service removes the operational burden but charges a premium; self-hosting on your own instances is cheaper per unit but demands real operational effort. Which wins depends on scale and how much of a platform team you have. Here is the comparison with numbers.
Self-managed Kafka, MSK, and Kinesis price streaming very differently: broker instances plus replication transfer versus shards or throughput units. The cheapest choice depends on throughput, retention, and how much operational work you want. Here is the breakdown.
Streaming processes data continuously for low-latency freshness, keeping compute always running; batch processes in scheduled bulk runs on transient compute. When data does not need to be fresh instantly, batch is far cheaper. Here is the tradeoff.
Kafka cost is the broker compute, the storage of retained messages, cross-AZ replication transfer, and (if managed) the service premium. Retention, right-sized brokers, and tiered storage cut it. Here is how.
Stream Analytics bills per streaming unit-hour, the provisioned capacity that runs your streaming query, whether or not data flows. Right-sizing streaming units to real throughput and stopping idle jobs are the levers. Here is the model.
Event Hubs Standard bills throughput units by the hour plus per-million ingress events; Premium and Dedicated allocate reserved capacity. Under-provision throughput units and you throttle; over-provision and you overpay. Here is how to size it.
Firehose charges about $0.029 per GB ingested (first 500 TB/month), with extra for format conversion and VPC delivery. Cost tracks data volume, so batching and compression before ingest are the main levers. Here is the breakdown.
Kinesis scales down to a single shard with no real floor; MSK runs brokers 24/7 ($400+/month for a 3-broker cluster). Here's the crossover by throughput and the Kafka-ecosystem angle.