Aerospike Achieves One Million Writes Per Second on Google Compute Engine with Just 50 Nodes

Aerospike is used as a front edge operational database for a variety of purposes: a session or user context store for real-time bidding, personalization, fraud detection, and real-time analytics. These applications must read and write billions of keys and terabytes, from click-streams to sensor data. Data in Aerospike is replicated synchronously in-memory to ensure immediate consistency and written to disk asynchronously. The combined Aerospike-Google Cloud Platform solution could fuel an entirely new category of applications that must process data in real-time and at scale from the very start, enabling a new class of startups with business models that were not viable economically previously.

Aerospike, an open-source, flash-optimized, in-memory NoSQL database, was looking to push the boundaries of Google's speed on Google Compute Engine. The challenge was to meet high throughput, consistently low latency, and real-time processing, which are characteristic of future cloud applications. The team at Aerospike was inspired by Ivan Santa Maria Filho, Performance Engineering Lead at Google, who demonstrated 1 Million Writes Per Second with Cassandra on Google Compute Engine. The goal was to benchmark Aerospike's product performance on Google Compute Engine and see if it could scale with consistently low latency, require smaller clusters, and be simpler to operate.

Aerospike conducted a benchmark using a similar setup as the Cassandra benchmark: 100 Million records at 200 bytes each, debian 7 backports, servers on n1-standard-8 instances with data-in-memory with on-disk persistence on a 500GB non-SSD persistent disks at $0.504/hr, clients on n1-highcpu-8 instances at $0.32/hr. They documented pure write performance, pure read, and mixed read/write performance. The results showed that Aerospike scaled on Google Compute Engine with consistently low latency, required smaller clusters, and was simpler to operate. The combined Aerospike-Google Cloud Platform solution could fuel an entirely new category of applications that must process data in real-time and at scale from the very start, enabling a new class of startups with business models that were not viable economically previously.

• The benchmarking of Aerospike on Google Compute Engine demonstrated high throughput for both reads and writes, consistent low latency, and unmatched price/performance. The results showed that Aerospike scaled on Google Compute Engine with consistently low latency, required smaller clusters, and was simpler to operate. This performance could enable a new generation of applications with mixed read/write data access patterns that sense and respond to what users do on websites and on mobile apps across the Internet in real-time. The combined Aerospike-Google Cloud Platform solution could fuel an entirely new category of applications that must process data in real-time and at scale from the very start, enabling a new class of startups with business models that were not viable economically previously.

• 1 Million Writes per Second with just 50 Aerospike servers

• 1 Million Reads per Second with just 10 Aerospike servers

• 7ms median latency for Writes with 83% of writes < 16ms and 96% < 32