SQLstream Blaze demonstrates its stream processing performance credentials using the industry-standard WordCount benchmark. For enterprises using Blaze, the results mean real-time Big Data scalability at a fraction of the cost of other solutions.
SAN FRANCISCO, CA – November 12, 2014 – SQLstream Inc., the provider of the real-time data hub for enterprise grade stream processing, today announced the results of a comparative performance benchmark between SQLstream Blaze and the open source stream processing framework, Apache Storm. Using the industry standard WordCount benchmark, a measure based on of the number words processed per second, SQLstream Blaze achieved a throughput of 4.68 million records per second per server against Apache Storm’s maximum throughput of 41,376 records per second per server. SQLstream Blaze performed 113X faster on equivalent hardware than Apache Storm.
The comparative benchmark was performed using Intel® Xeon® X5560 2.8GHz 8-core servers each with 32GB RAM. The WordCount benchmark is included as the WordCount Topology with the Apache Storm distribution, where a WordCount Bolt reads sentences as a stream of words from a Spout and outputs an updated count of the total number words or the total number of times each word has been received.
The SQLstream Blaze Stream Processor is a real-time data hub for operational intelligence (log data) and Internet of Things (sensor data) applications that is built on the award-winning distributed SQL stream processing platform, s-Server. SQLstream s-Server is the only stream processing platform that supports standards-compliant SQL queries for data stream processing and streaming analytics. Continuous queries utilize standard SQL operators such as SELECT, INSERT, JOIN (outer and inner), GROUP BY, PARTITION BY and WINDOW to process time-series data streams in real-time. SQL also enables streaming queries to be planned, optimized and distributed automatically over multiple cores and servers.
“Enterprises require assurances of platform scalability and performance but are increasingly concerned with the escalating hardware and overall solution costs of Hadoop and related stream processing solutions,” said Damian Black, “In all regards – performance, stability and total cost of ownership - this benchmark solidifies SQLstream’s position as the stream processor of choice for the real-time enterprise.”
Here is an example of a simple continuous SQL query from the WordCount benchmark, a tumbling window for streaming aggregation that reads from a “word” stream. The output is a new stream of records, one per second, specifying the total number of words received and processed in that second.
SELECT STREAM "word", SUM("count") AS "count" FROM "word_count" GROUP BY FLOOR("word_count".rowtime TO SECOND), "word";
Download Blaze for Free | Perform the Benchmark
Full details of the WordCount benchmark with Apache Storm and SQLstream Blaze, including the SQL code and access to the free download of Blaze, are all available on the SQLstream website at www.sqlstream.com/stream-processing-performance.
SQLstream solves real-time enterprise problems with unrivaled expertise, experience and success. Our SQLstream Blaze Stream Processor is a massively scalableaskoxford. real-time data hub for operational intelligence and Internet of Things applications that is built on our award-winning stream processing platform. Enterprises using Blaze act immediately on actionable insights extracted from their machine data in motion through real-time alerts and automated actions. Blaze eliminates the complexity, time and cost of real-time performance, and enables enterprises to deliver a better customer experience, increase revenue, improve operational efficiency and eliminate fraud. SQLstream is the recipient of leading industry awards, including the Ventana Research Technology Innovation Award for IT Analytics and Performance. SQLstream is based in San Francisco, California. For more information, visit www.sqlstream.com, follow us on Twitter, or visit sqlstream.com/stream-processing for a full tutorial on stream processing with SQL.