Pepperdata announced
that the Pepperdata product portfolio now includes the ability to monitor
Graphics Processing Units (GPUs) running big data applications like Spark on
Kubernetes.
Workloads that harness tremendous amounts of data, such as machine learning
(ML) and artificial intelligence (AI) applications, require GPUs, which were
originally designed to accelerate graphics rendering. That extra processing
power comes with a high price tag, and it requires near constant monitoring for
resource waste to get the best performance at the lowest possible cost.
Pepperdata now monitors GPU performance, providing the visibility needed for
Spark applications running on Kubernetes and utilizing the processing power of
GPUs. With this new visibility, companies can improve the performance of their
Spark apps running on those GPUs and manage costs at a granular level.
Unlike traditional infrastructure monitoring, which is limited to the
platform, the Pepperdata solution provides visibility into GPU resource utilization
at the application level. Pepperdata also provides instant recommendations for
optimization. Features include:
- Visibility
into GPU memory usage and waste
- Fine-tuning
of GPU usage through end-user recommendations
- Ability
to attribute usage and cost to specific end-users
"Spark on Kubernetes is quickly becoming a dominant part of the compute
infrastructure as data-intensive ML and AI applications proliferate," said
Ash Munshi, CEO, Pepperdata. "GPUs can handle these workloads, but they
are expensive to buy and are power-intensive. Until now, there hasn't been a
way to view and manage the infrastructure and applications, which can lead to
unnecessary waste and overspending for big data workloads. With Pepperdata,
organizations can properly size their GPU hardware investments and have the
confidence that they are utilizing them well."
There are products on the market for monitoring GPUs, but they typically
lack long-term storage, the ability to scale, and often do not correlate
infrastructure metrics to applications. Pepperdata solves these problems with
insight for data center operators, data scientists, and ML/AI developers. They
can now understand who is using what resources, optimize to eliminate waste so
jobs can be tuned and prioritized, and make sure costs are assigned
appropriately to the right users or groups across the enterprise.