HPE Nonstop: Enhancing openness & new scale-up ability for mission-critical workloads

[ComputeExperts]

HPE has modernized its HPE Nonstop threading with kernel-level threads (KLT), helping applications scale easily and enabling faster porting from open platforms.

With kernel-level threading, easily scale up and scale out your workloads with the industry’s best-in-class availability

 HPE’s introduction of kernel-level threads (KLT) on the HPE Nonstop platform marks a strategic modernization of its threading architecture, realizing a compute platform that can both scale up and scale out at the same time—a unique feat in the industry.

What’s changing?

Historically, HPE Nonstop applications relied on cooperative threading models (PUT, SPT, GTHREADs), where multiple user threads shared a single kernel thread. This limited scalability as it restricted threads of a process to a single CPU core. With KLT, each user thread is backed by a kernel thread, enabling true parallel processing across CPU cores. Support for POSIX application programming interfaces (APIs) and services such as mutexes, read-write locks, thread local storage, and others offer the ability for applications built for platforms such as Linux and Windows to be easily ported to run on the HPE Nonstop platform, allowing these workloads to gain the high-availability (HA) advantages out of the box. Let’s look at a before-and-after scenario to understand this better:

Before KLT

HPE Nonstop solutions offered a scale-out architecture, unique in the industry. However, applications on multicore processors had to use a multiprocess architecture rather than a multithreaded architecture to be able to run on all cores i.e., scale up. Hence, multithreaded applications had to be modified to achieve outstanding scalability.

After KLT

With the new implementation, multithreaded applications can now seamlessly take advantage of the multiple cores within an HPE Nonstop system without being re-architected. They can now achieve both scale-out and scale-up capabilities—a unique characteristic among the computing platforms available today. Hence, applications developed on HPE Nonstop can scale as no other, and those supported on Linux platforms can easily be ported to HPE Nonstop and achieve best-in-class scalability out of the box.

Why does it matter for business?

• Scale up: For applications currently running on HPE Nonstop, harness multicore performance for real-time workloads. Gain scale-up advantages by utilizing all cores of an HPE Nonstop CPU in addition to scale-out processing across multiple CPUs of a system.
• Portability: Multithreaded applications developed for Linux or Windows can be more easily ported to HPE Nonstop, reducing development overhead and accelerating time to market. Maintain a common code base for your application across different supported platforms and reduce development costs.
• Operational efficiency: Better CPU utilization means lower infrastructure costs and improved responsiveness for mission-critical systems.

 

Strategic implications

HPE Nonstop is a modern, standards-aligned platform for mission-critical enterprise workloads. It enables:

• Workload modernization offering improved support for modern application environments based on Java, Python, and others
• Development of solutions using design concepts, including microservices, polyglot programming, and DevSecOps principles on scalable and highly available platforms
• Deployment of real-time transaction processing workloads with reduced latency
• Easier adoption of open-source and cross-platform software

For organizations invested in HPE Nonstop, KLT is not just a technical upgrade—it’s a strategic enabler for hybrid architectures, modernization initiatives, and long-term platform viability. For organizations trying to break free from costly vendor lock-ins, looking to improve their application resilience, and move beyond on-paper high-availability vendor claims, this capability offers an easy path to port their applications onto a trusted platform and gain the mission-critical advantage out of the box. The HPE Nonstop platform is now also available with GreenLake, the cloud that comes to you, making it all the easier to adopt this technology in a cloud economics model.

For more information, refer to the technical brief and HPE Nonstop Threading User Guide.

Learn more at: hpe.com/info/Nonstop

By Author:
Prashanth Kamath
Product Manager,
HPE Nonstop Enterprise Division (NED)