Altair has stated that it acquired Ellexus– an input/output (I/O) profiling firm, which helps companies understand the way they access data for various applications. The acquisition of Ellexus is said to provide Altair with more storage functionality for high-performing computer solutions across industries.
“Altair proceeds to extend its reach and capacities for HPC environments to support critical modern workloads including for data analytics, AI and advanced driver-assistance systems (ADAS),” said James Scapa, Altair’s chief executive officer and founder. “The acquisition of Ellexus is especially relevant in these domains as storage-aware scheduling for big data applications is crucial.”
Altair will also deepen its technical skillset as Dr Rosemary Francis, Ellexus founder and the chief executive officer will join its team to help Altair best utilise the technology. “There is no better place for Ellexus’ solutions to get into the hands of global customers who can immediately benefit,” said Francis. “Altair’s expanding leadership in HPC is exciting and precisely where I want to be able to help advance the business and expand the technology to its limits.”
Why Is I/O Analysis Important For High-Performance Computing?
A large expenditure in the cloud can be wasted on I/O intensive applications, particularly when it comes to storage needs. According to experts, it’s very easy to over-provision your storage needs in the cloud and that’s where a lot of the costs can add up. Understanding your needs and making sure that you’re not over-provisioning or under-provisioning is really the key.
I/O profiling is used to help solve performance issues, both in systems and applications. It is also used to prepare and protect shared storage from applications that might be overloading shared resources. It’s also used to look into other issues such as managing application dependencies and making sure which files have been accessed, where they’re being stored and so on. More recently within the HPC community, I/O profiling is used heavily to manage hybrid cloud resources.
Detecting dependencies is another thing that you want to do with I/O profiling. Here, companies want to analyse the file path, the package that was used to install a file, the file name, the file type and the mount point. In fact, dependencies within some industries are extremely crucial, like EDA and HPC. Customers run products such as breeze-as-a-service where users submit their job to a particular queue, and they get back an automated list of dependencies in order to check the correctness of their application for containerisation, migration and so on.
Expanding HPC Capabilities
Ellexus products — Mistral and Breeze — are utilised for I/O diagnostics, optimisation and dependency analysis by HPC administrators of big companies. The power of the cloud can only be utilised if companies have good visibility into applications, both on-prem and in the cloud. When enterprises deploy HPC and other scientific applications to the cloud, they actively require the capacity to understand I/O patterns, resource use and system performance to tune and maintain data workloads seamlessly.
This will help optimise throughput and performance of apps, containers, and services to support clients and help manage thousands of applications, and run billions of tasks each day for real-time insights in HPC services. HPC is, in fact, emerging as an important element of AI transformation as it plays a crucial role in all fields of computational science and data analytics, particularly in the enterprise space.
With the transition to hybrid cloud and multi-cloud, many hardware vendors are innovating software to advance in the changing landscape. Here, system monitoring is growing to become more significant as the complexity of the HPC platform augments. System monitoring simplifies the migration of enterprise HPC workloads to the cloud while reducing costs as it provides real-time insights into workloads, and spending with complete visibility to HPC cloud resources.
Altair’s Strategy
Post the acquisition; it has been reported that these products will be combined into the storage aware scheduling functionality of Altair PBS Works. This will help the company improve its HPC prowess, thus solving much of the dependencies issues of its clients.
With analytics, software product development and HPC services, Altair is working alongside companies across large industry segments, all the way from oil and gas to finance giants. The addition of Ellexus will bring in Altair the capability of implementing its best practices in terms of I/O performance in HPC.
Altair has been one of the leading providers of HPC workload and workflow management tools for sectors like manufacturing, research, oil and gas, weather, and government. It acquired another company, Univa, which provides enterprise-grade workload management, scheduling, and optimisation products for HPC and AI/ML workloads both on-prem and in the cloud.
The acquisition of Univa and now Ellexus helps Altair position itself as a leading scheduling and resource optimisation solution provider for both massively parallel and high-throughput, single-core tasks. Acquiring Ellexus also complements Altair’s scheduling capabilities and network I/O real-time monitoring, ensuring quicker job execution and greater resource utilisation.
These acquisitions will further improve Altair’s capability and performance needs for its clients and strengthen the company’s position in workload management and cloud enablement for high-performance computing.
