Improving Application Performance with NVMe Storage - Part 1
The Rise of AI and ML Driving Parallel Computing Requirements
April 29, 2019

Zivan Ori
E8 Storage

Share this

As computing technology and data algorithms have advanced over the years, the ways in which technology has been applied to real world challenges have grown more automated and autonomous. This has given rise to a completely new set of computing workloads for Machine Learning which drives Artificial Intelligence applications (aka AI / ML).

AI / ML can be applied across a broad spectrum of applications and industries. Financial analysis with real-time analytics is used for predicting investments and drives the FinTech industrys needs for high performance computing. Real-time image recognition is a key enabler for self-driving vehicles, while facial recognition is used by law enforcement across the globe. Manufacturing uses image recognition technology to spot defects in materials, organizations such as NOAA use satellite imagery to spot changes in weather, while social media platforms use image recognition to tag photos of friends and family.

What is common among these uses cases is the need for a high level of parallel computing power, coupled with a high-performance low latency architecture to enable parallel processing of data in real-time across the compute cluster. The "training" phase of machine learning is critical and can take an excessively long time, especially as the training data set grows exponentially to enable deep learning for AI.

With storage performance now recognized as a critical component of AI/ML application performance, the next step is to identify the ideal storage platform. Non-Volatile Memory Express (NVMe) based storage systems have gained traction as the storage media of choice to deliver the best throughput and latency. Shared NVMe storage systems unlock the performance of NVMe, and offer a strong alternative to using local NVMe SSDs inside of GPU nodes.

The Rise of GPUs for AI / ML

GPUs were originally created for high performance image creation, and are very efficient at manipulating computer graphics and image processing. Their highly parallel structure makes them much more efficient than general purpose CPUs for algorithms where the processing of large blocks is done in parallel. For this reason, GPUs have found strong adoption in the AI / ML use case as they allow for a high degree of parallel computing and current AI focused applications have been optimized to run on GPU based computing clusters.

With the powerful compute performance of GPUs, the bottleneck moves to other areas of the AI / ML architecture. For example, the volume of data required to feed machine learning requires massive parallel read access to shared files from the storage subsystem across all nodes in the GPU cluster. This creates a performance challenge that NVMe shared storage systems are ideally suited to address.

Shared NVMe Storage for High Performance Machine Learning (ML)

One of benefits of shared NVMe storage is the ability to create even deeper neural networks due to the inherent high performance of shared storage, opening the door for future models that cannot be achieved today with non-shared NVMe storage solutions.

Today, there are storage solutions that offer patented architectures built from the ground up to leverage NVMe. The key to performance and scalability is the separation of control and data path operations between the the storage controller software and the host-side agents. The storage controller software provides centralized control and management, while the agents manage data path operations with direct access to shared storage volumes.

While AI / ML workloads are run exclusively on the GPUs within the cluster, that doesn't mean that CPUs have been eliminated from the GPU clusters completely. The operating system and drivers still leverage the CPUs, but while the machine learning training is in progress, the CPU is relatively idle. This provides the perfect opportunity for an NVMe based storage architecture to leverage the idle CPU computing capacity for a high performance distributed storage approach.

With NVMe protocol supporting exponentially more connections per SSD, the storage agents use RDMA to give each GPU node a direct connection to the drives. This approach enables the agents to perform up to 90% of the data path operations between the GPU nodes and storage, reducing latency to be on par with local SSDs.

In this scenario, running the NVMe based storage agent on the idle CPU cores of the GPU nodes enables the NVMe based storage to deliver 10x better performance than competing all-flash solutions, while leveraging existing compute resources that are already installed and available to use.

Read Part 2: Local versus Shared Storage for Artificial Intelligence (AI) and Machine Learning (ML)

Zivan Ori is CEO and Co-Founder of E8 Storage
Share this

The Latest

June 25, 2020

I've had the opportunity to work with a number of organizations embarking on their AIOps journey. I always advise them to start by evaluating their needs and the possibilities AIOps can bring to them through five different levels of AIOps maturity. This is a strategic approach that allows enterprises to achieve complete automation for long-term success ...

June 24, 2020

Sumo Logic recently commissioned an independent market research study to understand the industry momentum behind continuous intelligence — and the necessity for digital organizations to embrace a cloud-native, real-time continuous intelligence platform to support the speed and agility of business for faster decision-making, optimizing security, driving new innovation and delivering world-class customer experiences. Some of the key findings include ...

June 23, 2020

When it comes to viruses, it's typically those of the computer/digital variety that IT is concerned about. But with the ongoing pandemic, IT operations teams are on the hook to maintain business functions in the midst of rapid and massive change. One of the biggest challenges for businesses is the shift to remote work at scale. Ensuring that they can continue to provide products and services — and satisfy their customers — against this backdrop is challenging for many ...

June 22, 2020

Teams tasked with developing and delivering software are under pressure to balance the business imperative for speed with high customer expectations for quality. In the course of trying to achieve this balance, engineering organizations rely on a variety of tools, techniques and processes. The 2020 State of Software Quality report provides a snapshot of the key challenges organizations encounter when it comes to delivering quality software at speed, as well as how they are approaching these hurdles. This blog introduces its key findings ...

June 18, 2020

For IT teams, run-the-business, commodity areas such as employee help desks, device support and communication platforms are regularly placed in the crosshairs for cost takeout, but these areas are also highly visible to employees. Organizations can improve employee satisfaction and business performance by building unified functions that are measured by employee experience rather than price. This approach will ultimately fund transformation, as well as increase productivity and innovation ...

June 17, 2020

In the agile DevOps framework, there is a vital piece missing; something that previous approaches to application development did well, but has since fallen by the wayside. That is, the post-delivery portion of the toolchain. Without continuous cloud optimization, the CI/CD toolchain still produces massive inefficiencies and overspend ...

June 16, 2020

The COVID-19 pandemic has exponentially accelerated digital transformation projects. To better understand where IT professionals are turning for help, we analyzed the online behaviors of IT decision-makers. Our research found an increase in demand for resources related to APM, microservices and dependence on cloud services ...

June 15, 2020

The rush to the public cloud has now slowed as organizations realized that it is not a "one size fits all" solution. The main issue is the lack of deep visibility into the performance of applications provided by the host. Our own research has recently revealed that 32% of public cloud resources are currently under-utilized, and without proper direction and guidance, this will remain the case ...

June 11, 2020

The global shift to working from home (WFH) enforced by COVID-19 stay-at-home orders has had a massive impact on everyone's working lives, not just in the way they remotely interact with their teams and IT systems, but also in how they spend their working days. With both governments and businesses committed to slowly opening up offices, it's increasingly clear that a high prevalence of remote work will continue throughout 2020 and beyond. This situation begets important questions ...

June 10, 2020
In recent years, with the emergence of newer technologies ranging from the cloud to machine learning, IT modernization has evolved from a replacement of end-of-life infrastructure to an enabler of innovation and business value. It is a complex process that can take months or even years, but a recent survey shows that the effort begins to deliver measurable results almost as soon as an organization executes the first steps on its roadmap ...