OpenStack is open-source software which is used for building public or private clouds. Competitors include VMware, vCloud and Amazon Web Services (News - Alert), so steps need to be taken to make OpenStack as popular as possible in the competitive marketplace.
Customers also need to know how well OpenStack performs, but OpenStack is complex with many subsystems and components, which makes it hard for customers or potential customers to predict how “different implementation decisions and changes proposed to OpenStack affect the whole system’s behavior and performance,” according to a statement from Mirantis. On top of this, there are many alternative cloud configurations where OpenStack can be installed and used. Benchmarking becomes very difficult because it’s hard to get results from just parts of OpenStack. What was needed is a way to find benchmarking on the entire OpenStack system, with a predefined cloud configuration, the company explained. The benchmarking tool would also need to come up with complex and reproducible scenarios on actual OpenStack deployments.
In other words, customers basically need to monitor how well of a job OpenStack is doing. So IBM, SoftLayer (News - Alert) (part of IBM) and Mirantis have developed Rally. It’s a benchmarking tool that reports on OpenStack performance.
“To accelerate OpenStack adoption among mainstream enterprises, it is essential to provide customer confidence in the product by establishing credible performance benchmarks,” Adrian Ionel, Mirantis’ CEO, said in a recent statement.
In a recent interview, Ionel also told GigaOM some potential large customers wanted more data on OpenStack performance. “We had estimates but not hard data on how it performs on a variety of loads and configurations, and performance can be configuration specific. The tooling in the OpenStack framework did not exist for this so we built it and contributed it back to the community.”
Before, there were tools that could be used in benchmarking for OpenStack use. Examples are DevStack and FUEL, which deploy and manage OpenStack clouds, Tempest (News - Alert), which validates OpenStack APIs and Tomograph, which is used for tracing.
“The challenge, however, is to combine all these tools together on a reproducible basis,” an OpenStack document said. That’s where Rally comes into play.
“Compiling a bunch of already existing benchmarking facilities into one project, making it flexible to user requirements and ensuring the reproducibility of test results, is exactly what Rally does,” the statement explained. Rally provides many metrics on performance thanks in part to the deployment of OpenStack on over 1,500 bare-metal servers from SoftLayer.
Rally sends its reports to a database, and provides data profiles. It can find performance and scale problems and explain why a cloud may be slower. Consider the specific wide variety of uses that Rally can provide. OpenStack listed several of these. It can help show how deployments can impact the performance of the operating system. It will provide deployment specs for different kinds of loads. (These may include controllers, swift nodes, and others.) It can also search to find the best hardware for a specific OpenStack cloud and monitor the performance of cloud operations. Or, it can automate monitoring on how code changes will impact the performance of the operating system.
By using Rally, Silicon Angle reports that users can “simulate applications with different workflows and behavior properties, optimize their hardware for high-performance and spot bottlenecks, as well as automate various administrative tasks.”
Some of its key elements are: a deploy engine, benchmarking tool, server provider and orchestrator.
“When using OpenStack for production workloads, enterprises can be confident that OpenStack performance will meet the SLAs specified for those workloads,” Sesh Murthy, vice president, Architecture and Technical Solution Design at IBM (News - Alert), stated. “What is great about Rally is that it allows you to run a simulated workload on OpenStack for a period of time and report on workload performance against the SLA. This provides much-needed statistical data on infrastructure performance prior to movement into production.”
So to remain competitive, OpenStack needs to provide this kind of benchmarking data to its customers. The results may lead to growth for both OpenStack and its enterprise customers.