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

May 23, 2019

The first word in APM technology is "Application" ... yet for mobile, apps are entirely different. As the mobile app ecosystem is evolving and expanding from pure entertainment to more utilitarian uses, there's a rising need for the next generation of APM technology to stay ahead of the issues that can cause apps to fail ...

May 22, 2019

For application performance monitoring (APM), many in IT tend to focus a significant amount of their time on the tool that performs the analysis. Unfortunately for them, the battle is won or lost at the data access level. If you don’t have the right data, you can’t fix the problem correctly ...

May 21, 2019

Findings of the Digital Employee Experience survey from VMware show correlation between enabling employees with a positive digital experience (i.e., device choice/flexibility, seamless access to apps, remote work capabilities) and an organization's competitive position, revenue growth and employee sentiment ...

May 20, 2019

In today's competitive landscape, businesses must have the ability and process in place to face new challenges and find ways to successfully tackle them in a proactive manner. For years, this has been placed on the shoulders of DevOps teams within IT departments. But, as automation takes over manual intervention to increase speed and efficiency, these teams are facing what we know as IT digitization. How has this changed the way companies function over the years, and what do we have to look forward to in the coming years? ...

May 16, 2019

Although the vast majority of IT organizations have implemented a broad variety of systems and tools to modernize, simplify and streamline data center operations, many are still burdened by inefficiencies, security risks and performance gaps in their IT infrastructure as well as the excessive time it takes to manage legacy infrastructure, according to the State of IT Transformation, a report from Datrium ...

May 15, 2019

When it comes to network visibility, there are a lot of discussions about packet broker technology and the various features these solutions provide to network architects and IT managers. Packet brokers allow organizations to aggregate the data required for a variety of monitoring solutions including network performance monitoring and diagnostic (NPMD) platforms and unified threat management (UTM) appliances. But, when it comes to ensuring these solutions provide the insights required by NetOps and security teams, IT can spend an exorbitant amount of time dealing with issues around adds, moves and changes. This can have a dramatic impact on budgets and tool availability. Why does this happen? ...

May 14, 2019

Data may be pouring into enterprises but IT professionals still find most of it stuck in siloed departments and weeks away from being able to drive any valued action. Coupled with the ongoing concerns over security responsiveness, IT teams have to push aside other important performance-oriented data in order to ensure security data, at least, gets prominent attention. A new survey by Ivanti shows the disconnect between enterprise departments struggling to improve operations like automation while being challenged with a siloed structure and a data onslaught ...

May 13, 2019

A subtle, deliberate shift has occurred within the software industry which, at present, only the most innovative organizations have seized upon for competitive advantage. Although primarily driven by Artificial Intelligence (AI), this transformation strikes at the core of the most pervasive IT resources including cloud computing and predictive analytics ...

May 09, 2019

When asked who is mandated with developing and delivering their organization's digital competencies, 51% of respondents say their IT departments have a leadership role. The critical question is whether IT departments are prepared to take on a leadership role in which collaborating with other functions and disseminating knowledge and digital performance data are requirements ...

May 08, 2019

The Economist Intelligence Unit just released a new study commissioned by Riverbed that explores nine digital competencies that help organizations improve their digital performance and, ultimately, achieve their objectives. Here's a brief summary of 7 key research findings you'll find covered in detail in the report ...