Grid and HPC

New challenges with HPC going mainstream

With commodity X64 hardware and interconnects flooding the market, HPC and distributed computing are going mainstream. Every day, more applications and more users benefit from this new technology paradigm. The workload scheduler is a more and more frequent component of clusters, and for many applications it effectively becomes the main Network Operating System.

These trends also generate new challenges:

  • deployment, monitoring and management of systems from different vendors
  • heterogeneity in memory size, CPU performance, network and disk I/O
  • different workload profiles for different applications
  • ad-hoc data distribution and access

If not properly addressed, these challenges quickly lead to the creation of incompatible, application-wrapped silos of clusters, which lead to a waste in resources, penalties in application performance by under-provisioning, and increased management overhead.

Our solutions

Leveraging With 12+ years of successes at commercial and academic leaders, NICE can deliver solutions and services to build robust, efficient Enterprise Grids. By combining market leading products with competent professional services, we simplify management and usage of heteroneneous, multi-application computing infrastructures ready to face the most challenging production workloads.

The key components of our solutions are:

  • Scalable, flexible workload scheduling software, including commercial or open source products from Adaptive Computing, IBM and Univa 
  • Professional Services and qualified, responsive Support services
  • Innovative R&D projects to pioneer new technologies from the Open Source and Research world

Coupled with our Grid Portal, Remote Visualization, Cloud and Grid Intelligence solutions, our customers enjoy great efficiency and productivity while deploying, using and managing complex computing infrastructures.

End users can typically enjoy the following improvements:

  • Better application performance and throughput by fine-tuned workload balancing and queuing
  • Better service levels driven by business policies
  • Increased service uptime and reliability
  • Additional computing capacity by tapping into unused capacity shared with other groups / applications

On the other side, the Enterprise Grid delivers significant added-value for the system administrators:

  • Full control and accounting of every computing activity
  • Centralized configuration and policy definition
  • Great scalability and flexibility to support any workload running on Linux, UNIX or Windows
  • SLA-based consolidation of resources from different groups

 (features may vary according to the choice of scheduler)