Advanced Computing in the Age of AI | Wednesday, June 19, 2024

VoltDB Database Put Through The Paces On Clouds 

One of the buzzwords for users of cloud services refers to the increasing need for speed and the velocity of data. Datasets are not only growing in size but for many applications must be manipulated in real time. Hence, it is argued, friction in the cloud must be reduced to handle accelerating data.

An example of the shift to "fast data" is the growing volumes of sensor information generated by the emerging Internet of Things. Depending on the application, real-time processing can be achieved through virtualization, more powerful hardware, or a combination of both.

That appears to be one of the issues being addressed by IBM partner and database vendor VoltDB, which has unveiled a NewSQL in-memory database the company claims runs applications much faster on IBM's SoftLayer cloud platform than on Amazon Web Services (AWS). Running the database on the SoftLayer infrastructure would deliver souped-up analytics on large sets of "fast moving data" to enable "real-time transactional decisions," the company claimed.

The company also touts VoltDB as a "ground-up redesign of the relational database" that can be scaled to handle growing volumes of fast data.

To support its claim that VoltDB runs applications up to five times faster on SoftLayer than on AWS, the privately held company provided performance benchmarks based on the Yahoo Cloud Serving Benchmark (YCSB). It tested the in-memory database on clusters with from three to a dozen clusters running on the SoftLayer platform, which IBM touts as best suited to "data-driven” applications.

A bare metal benchmark was run on commodity two-socket servers using Intel processors. VoltDB was directly installed after provisioning along with a Java virtual machine. Three main YCSB workloads were targeted, including a 50-50 percent mix of read and update operations.

For the 50-50 workload, throughput on SoftLayer was found to increase from 426,000 transactions per second for three nodes to 904,000 transactions per second at twelve nodes. VoltDB said that was 2.5 times faster than AWS.


The company also claimed its relational database running on SoftLayer delivered performance equivalent to bare metal along with "fast data processing capabilities" used for big data analytics. Ultimately, VoltDB helps "close the [big] data loop, the company asserts.

The tool also serves to broaden the SoftLayer cloud ecosystem for boosting the deployment speed and flexibility of cloud applications. IBM is promoting fast data applications on the cloud as a way of "unlocking the economic value" of big data through "real-time and automated decision making."

Fast data applications running in the cloud require support for options ranging from "multitenant virtualization to bare metal performance," Mac Devine, CTO of IBM's Cloud Services Division, said in a statement released by VoltDB.

Along with its in-memory database, VoltDB also announced it is joining IBM's Cloud Marketplace that combines IBM's capabilities-as-a-service offerings with those of partners and third-party vendors.

VoltDB has offices in Bedford, Massachusetts and Santa Clara, California.

About the author: George Leopold

George Leopold has written about science and technology for more than 30 years, focusing on electronics and aerospace technology. He previously served as executive editor of Electronic Engineering Times. Leopold is the author of "Calculated Risk: The Supersonic Life and Times of Gus Grissom" (Purdue University Press, 2016).