Are you stream processing records that contain IP addresses? If you process events or logs, perform security and threat analysis, or capture streaming events generated by IoT, you probably are.
The most common and most basic IP-based enrichment in these contexts is appending the IP address's corresponding hostname (the DNS PTR value) to each record. The trouble is, performing a DNS query to resolve a hostname (the standard method) can be painfully slow. So slow, it kills your throughput.
Applications can be designed to 'go wide' with parallelism. These recover throughput at the cost of increased complexity and operating costs. But no matter how parallel they are, if you use DNS the latency remains.
|The Old Way:||Start a DNS lookup. Wait (maybe forever).|
|A Better Way:||Fetch from our managed cache. Move ahead.|
Dedistributed offers a better way. We curate and keep the entire 50 GB dataset in RAM on AWS-resident servers. A purpose-specific cache design, per-AZ client-to-server matching, and a binary TCP protocol enable lightning-fast retrievals of any (or all) of the 1.3 billion public hostname records. Our caches return results to your applications with sub-millisecond latency, even when you request the hostnames of 1000 IP addresses at once.
Dedistributed PTR is for today's streaming and batch workloads. It is a simple, easy-to-use drop-in solution. Subscribe through the AWS Marketplace. Use in the cloud or on-premises. Our pricing reflects the commodity nature of the data itself.
In the AWS Marketplace
Our target customers use Kafka, Elastic, Kinesis, Spark, Samza, Storm, and Flink. Batch mode applications are supported too.