Redis Emerges As NoSQL In-Memory Performer

With multiple NoSQL system choices, it might be hard to display a distinguishing feature. However, Redis Labs is outperforming some competitors during a recent in-memory big data system test.

Charles Babcock, Editor at Large, Cloud

June 26, 2015

3 Min Read
<p align="left">(Image: Shivendu Jauhari/iStockphoto)</p>

COBOL Leads Us Back To The Future

COBOL Leads Us Back To The Future


COBOL Leads Us Back To The Future (Click image for larger view and slideshow.)

An infrequently noticed system in the crowded NoSQL startup field is starting to creep up on its better known competitors thanks, in part, to its high-performance ratings. Redis Labs, the Israeli maker of Redis Labs Enterprise Cluster, shows up as the market performance leader in an Avalon Consulting LLC benchmark.

That might be because Redis is an in-memory system being matched up against the more typical memory and disk systems, such as MongoDB, Cassandra Open Source, and Couchbase. Two other NoSQL in-memory systems, however, were included in the test: Aerospike and Cassandra Datastax Enterprise In Memory. These are both in-memory systems, and Redis topped them as well, though the margin shrank somewhat.

All the vendors voluntarily participated in the benchmark after being solicited by Avalon to provide it with directions on how to optimize their systems for a benchmark test.

The test may or may not have been one that Redis was particularly well suited for, since it's Redis that is now handing out the results. Nevertheless, it offers a rare glimpse of competitors' performance with a publicly-defined application outlined for the test. In addition, in conducting the benchmark, Avalon solicited cooperation on the basis of the neutrality in how it conducted its previous test of Couchbase versus Cassandra in March.

Redis made the results known as it aired the fact that it has received an additional $15 million in funding to bring its system to market on June 25. The B round was led by the well-known Bain Capital Ventures and also Carmel Ventures, an Israeli VC group, bringing Redis' funding total to $28 million.

[Want to learn more about the crop of new NoSQL systems? See 16 NoSQL, NewSQL Databases To Watch.]

Redis now claims 4,900 paying customers and 30,000-plus free customers. Redis CEO Ofer Bengal said Redis Labs Enterprise Cluster is intended primarily for on-premises installation, but Redis also offers it as a managed service on AWS or as a service on IBM's Bluemix platform-as-a-service.

In its initial offering, Redis functions were entirely served out of server RAM on the cluster, but over the last two years it has engineered use of Flash memory attached to the server as an extension of RAM. At the moment, it's available in that form only on IBM Power 8 servers with Flash arrays, known as an IBM Big Data Engine, Bengal said.

"Redis is unique among NoSQL databases in that it is an in-memory NoSQL database which runs 1,000 times faster than disk databases," said Salil Deshpande, managing director at Bain Capital Ventures, in his blog announcing the Redis investment.

The Avalon benchmark described an application that is conducting a survey on the Web that at times is heavily trafficked and other times only lighted in use. That is, it's simulating an app that must rapidly scale out to high volumes, then scale back, based on its 2-CPU core database engine.

The benchmark found that Redis was able to process 37,039 application requests per second; Cassandra as an open source system processed 5,235; Cassandra as a Datastax in-memory system, 6,659; and Aerospike (also in-memory), 16,482.

In terms of the time it took to process each request, Redis had a latency of 71 milliseconds; Couchbase, 394; Cassandra open source, 381; Datastax Cassandra, 372; and Aerospike, 162.

"We have seven times the throughput of Couchbase, Cassandra, or Datastax Cassandra," said Bengal.

Redis was founded four years ago, with its database guru Salvatore Sanfilippo based in Sicily leading the open source Redis project. Redis has marketing and management offices in Mountain View and a technical staff in Tel Aviv.

About the Author

Charles Babcock

Editor at Large, Cloud

Charles Babcock is an editor-at-large for InformationWeek and author of Management Strategies for the Cloud Revolution, a McGraw-Hill book. He is the former editor-in-chief of Digital News, former software editor of Computerworld and former technology editor of Interactive Week. He is a graduate of Syracuse University where he obtained a bachelor's degree in journalism. He joined the publication in 2003.

Never Miss a Beat: Get a snapshot of the issues affecting the IT industry straight to your inbox.

You May Also Like


More Insights