Intel Labs Fights Silent Data Corruption with Computational Storage

By Linda Barney

May 30, 2024

Artificial Intelligence (AI) Large Language Models and other forms of Deep Learning already require enormous amounts of training data. The data volumes are expected to grow as more organizations implement AI. This situation introduces new system bottlenecks (compute, memory, and I/O) and raises new concerns about the real risks of silent data corruption. Organizations don’t want to question the integrity of their training data when building AI models.

Performing data integrity checks has traditionally been performed by CPUs, but these checks take time. It’s best to perform these checks continuously and in the background, a process known as data scrubbing, to minimize application disruption. Ideally, corrupted data is repaired before an application needs it. However, in practice, these integrity checks are often run infrequently or, worse, disabled entirely due to the cost. 

To address the growing concern of silent data corruption, Intel Labs is actively involved in research to solve the data integrity problem using computational storage (CS). According to Michael Mesnier, Intel Labs Principal Engineer, “Data scrubbing offload is a practical use case for computational storage, as it helps reduce data movement bottlenecks and related pressure on compute, memory, and I/O. Most storage devices already have built-in accelerators for checking data integrity, including checksum algorithms like CRC-64, but these accelerators have no file awareness. They only see blocks. Performing end-to-end file checksums within these devices requires computational storage techniques.”

How Can Computational Storage Assist with Data Integrity Checks?

Reducing I/O and associated processing is exactly the motivation behind computational storage, so data scrubbing represents a real use case. Checking the integrity of one file involves:

  • Reading the file.
  • Calculating a checksum (or hash) of the file.
  • Comparing that against a previously stored checksum.

If the checksums agree, the file’s integrity has not been compromised. If the checksums disagree, the file experienced corruption since it was last written, and it must be reconstructed (e.g., using replication or erasure coding).

And organizations must complete this process over their entire data set. It’s a read-intensive workload where you read every file, fromevery file system, on every drive in your storage cluster. If left unchecked, it’s an application killer, as it can consume all available I/O and increase load on host CPUs and memory.

Mesnier indicates that this is where computational storage can assist. By offloading data integrity checks to block storage, the checks can be performed in situ (inside a single SSD or storage server), and the costs associated with processing the I/O and running the integrity checks on the host CPU can be saved. But this must be done end-to-end. End-to-end file checksums are critical for data integrity, as silent corruption can occur in software (application, OS, device drivers), the CPU, memory, or I/O layer before data ever reaches the drive. The best practice is to calculate checksums as high in the stack as possible, as new data is being created, and this is usually done at the file level.

As Mesnier indicated, storage servers and SSDs already have a variety of built-in engines for integrity checks, but these are only performed at the block level, not end-to-end over an entire file. As such, offloading the integrity check will first require that we teach block storage about “files” and how to perform an end-to-end file checksum, which may be scattered across many regions of a storage device. Without computational storage techniques, a storage device cannot directly process files for a host, as it lacks data awareness. A storage device only sees sectors (hard drives) or pages (SSDs), not files and directories.

Intel Labs Computational Storage Research Platform

Researchers at Intel Labs created a research platform (currently available for customers and partners) to teach block storage how to see host data structures, like files, and subsequently process the data. “This has allowed us to tackle the challenges of computational storage and explore various use cases. Our research platform, which is based on the NVMe protocol, moves compute functions to either a storage server (e.g., NVMe/TCP) or a single SSD. This approach aligns well with industry-standard programming models that provide Computational Storage Functions (CSFs) via a Computational Storage Array (CSA) or Computational SSD (CSD). In both cases, offloading work to storage can reduce the host’s CPU load, memory footprint, I/O, and, in the case of a CSA, network traffic,” states Mesnier.  

As first described in their Hot Storage paper, the approach is based on virtual objects. In effect, virtual objects allow a host to share file metadata (e.g., the block mapping and the file size) with storage. This approach allows storage to “see” the file and read it into storage-local memory for subsequent processing. Virtual objects are embedded in compute descriptors along with a list of operations to be performed (e.g., search, filter, hash, checksum). These compute descriptors are sent to storage using new NVMe commands.

Various software layers build atop the virtual object mechanism, as shown in the figure below.

Computation storage software layers build atop the virtual object mechanism. (Source: Intel)

The application layer provides a convenient file-based interface and abstracts away most computational storage details. Applications simply request that operations be performed on files (e.g., search, filter, hash, checksum). This same layer can interface with other computing layers, like Intel’s oneAPI and FaaS.

The scheduling layer is for optimal scheduling across resources. The aggregation layer helps in processing data that is spread across multiple storage devices, which is the case with erasure-coded data. Finally, the device layer communicates with NVMe by creating computational storage commands and interfaces directly with CSAs and CSDs.

In the case of a CSA, Intel Labs has an additional layer that virtualizes and manages compute resources. This layer provides the foundation for a secure, multi-tenant programming environment. Within this block storage server, various silicon and systems software are used to accelerate compute and I/O-intensive operations. The solution uses Intel® Xeon®processors with support for CRC-64 (via the Data Streaming Accelerator) – a powerful end-to-end check for corrupted data.

Intel Labs has been using this platform both for research and for ecosystem enabling.

Building An Ecosystem

Computational storage will require strong ecosystem support, and Intel Labs is actively working with IHVs (Independent Hardware Vendors)and ISVs (Independent Software Vendors). IHV engagement is at the bottom of the stack (the device layer), and ISV engagement is at the top (the application layer).

This task involves working closely with SSD vendors, including Solidigm, to build CSD prototypes optimized for Intel server platforms. Scott Shadley (Director, Long-Term Strategy) at Solidigm indicates, “Solidigm’s PoC CSD is based on a Gen-5 SSD product, which includes a low-cost, high-efficiency and high-performance ASIC for data integrity calculations. We look forward to working with Intel to optimize the solution for server platforms and integrating it into a complete E2E software stack.”

At the top of the stack, Intel Labs has been working with MinIO to enable their stack for computational storage, beginning with data scrubbing offload. Early demonstrations of this collaborative work were presented last year at the Flash Memory Summit and SNIA’s Storage Developers Conference

Mesnier indicates that Intel Labs also aligns with SNIA to develop and evaluate computational storage usage models. SSD vendors can use their NVMe-based stack for prototyping and eventual standardization via TP4091.

Intel Labs Future Computational Storage Work

Computational storage has had many false starts going back decades, and the industry still lacks a widespread use case. According to Mesnier, “We believe that data scrubbing can be such a use case. The enormous data growth fueled by AI will make offloaded data integrity checks required, not just nice to have.” With a widespread use case in place, Intel Labs believes that the ecosystem will mature and become a foundation for other use cases. The same computational storage protocol used for data scrubbing can be higher up in the stack to accelerate other operations like AI (training and inference) and big data (sorting, searching, filtering). Intel Labs looks forward to the opportunities that computational storage will bring and is actively inviting others to collaborate.


Linda Barney is the founder and owner of Barney and Associates, a technical / marketing writing, training, and Web design firm in Beaverton, Oregon. The firm provides writing, training, and Web content for the high-tech, government, biotechnology, energy, medical, sustainability, high-performance computing, and scientific communities. 

Subscribe to HPCwire's Weekly Update!

Be the most informed person in the room! Stay ahead of the tech trends with industry updates delivered to you every week!

HPE and NVIDIA Join Forces and Plan Conquest of Enterprise AI Frontier

June 20, 2024

The HPE Discover 2024 conference is currently in full swing, and the keynote address from Hewlett-Packard Enterprise (HPE) CEO Antonio Neri on Tuesday, June 18, was an unforgettable event. Other than being the first busi Read more…

Slide Shows Samsung May be Developing a RISC-V CPU for In-memory AI Chip

June 19, 2024

Samsung may have unintentionally revealed its intent to develop a RISC-V CPU, which a presentation slide showed may be used in an AI chip. The company plans to release an AI accelerator with heavy in-memory processing, b Read more…

ASC24 Student Cluster Competition: Who Won and Why?

June 18, 2024

As is our tradition, we’re going to take a detailed look back at the recently concluded the ASC24 Student Cluster Competition (Asia Supercomputer Community) to see not only who won the various awards, but to figure out Read more…

Qubits 2024: D-Wave’s Steady March to Quantum Success

June 18, 2024

In his opening keynote at D-Wave’s annual Qubits 2024 user meeting, being held in Boston, yesterday and today, CEO Alan Baratz again made the compelling pitch that D-Wave’s brand of analog quantum computing (quantum Read more…

Apple Using Google Cloud Infrastructure to Train and Serve AI

June 18, 2024

Apple has built a new AI infrastructure to deliver AI features introduced in its devices and is utilizing resources available in Google's cloud infrastructure.  Apple's new AI backend includes: A homegrown foun Read more…

Argonne’s Rick Stevens on Energy, AI, and a New Kind of Science

June 17, 2024

The world is currently experiencing two of the largest societal upheavals since the beginning of the Industrial Revolution. One is the rapid improvement and implementation of artificial intelligence (AI) tools, while the Read more…

HPE and NVIDIA Join Forces and Plan Conquest of Enterprise AI Frontier

June 20, 2024

The HPE Discover 2024 conference is currently in full swing, and the keynote address from Hewlett-Packard Enterprise (HPE) CEO Antonio Neri on Tuesday, June 18, Read more…

Slide Shows Samsung May be Developing a RISC-V CPU for In-memory AI Chip

June 19, 2024

Samsung may have unintentionally revealed its intent to develop a RISC-V CPU, which a presentation slide showed may be used in an AI chip. The company plans to Read more…

Qubits 2024: D-Wave’s Steady March to Quantum Success

June 18, 2024

In his opening keynote at D-Wave’s annual Qubits 2024 user meeting, being held in Boston, yesterday and today, CEO Alan Baratz again made the compelling pitch Read more…

Shutterstock_666139696

Argonne’s Rick Stevens on Energy, AI, and a New Kind of Science

June 17, 2024

The world is currently experiencing two of the largest societal upheavals since the beginning of the Industrial Revolution. One is the rapid improvement and imp Read more…

Under The Wire: Nearly HPC News (June 13, 2024)

June 13, 2024

As managing editor of the major global HPC news source, the term "news fire hose" is often mentioned. The analogy is quite correct. In any given week, there are Read more…

Labs Keep Supercomputers Alive for Ten Years as Vendors Pull Support Early

June 12, 2024

Laboratories are running supercomputers for much longer, beyond the typical lifespan, as vendors prematurely deprecate the hardware and stop providing support. Read more…

MLPerf Training 4.0 – Nvidia Still King; Power and LLM Fine Tuning Added

June 12, 2024

There are really two stories packaged in the most recent MLPerf  Training 4.0 results, released today. The first, of course, is the results. Nvidia (currently Read more…

Highlights from GlobusWorld 2024: The Conference for Reimagining Research IT

June 11, 2024

The Globus user conference, now in its 22nd year, brought together over 180 researchers, system administrators, developers, and IT leaders from 55 top research Read more…

Atos Outlines Plans to Get Acquired, and a Path Forward

May 21, 2024

Atos – via its subsidiary Eviden – is the second major supercomputer maker outside of HPE, while others have largely dropped out. The lack of integrators and Atos' financial turmoil have the HPC market worried. If Atos goes under, HPE will be the only major option for building large-scale systems. Read more…

Comparing NVIDIA A100 and NVIDIA L40S: Which GPU is Ideal for AI and Graphics-Intensive Workloads?

October 30, 2023

With long lead times for the NVIDIA H100 and A100 GPUs, many organizations are looking at the new NVIDIA L40S GPU, which it’s a new GPU optimized for AI and g Read more…

Nvidia H100: Are 550,000 GPUs Enough for This Year?

August 17, 2023

The GPU Squeeze continues to place a premium on Nvidia H100 GPUs. In a recent Financial Times article, Nvidia reports that it expects to ship 550,000 of its lat Read more…

Everyone Except Nvidia Forms Ultra Accelerator Link (UALink) Consortium

May 30, 2024

Consider the GPU. An island of SIMD greatness that makes light work of matrix math. Originally designed to rapidly paint dots on a computer monitor, it was then Read more…

Nvidia’s New Blackwell GPU Can Train AI Models with Trillions of Parameters

March 18, 2024

Nvidia's latest and fastest GPU, codenamed Blackwell, is here and will underpin the company's AI plans this year. The chip offers performance improvements from Read more…

Choosing the Right GPU for LLM Inference and Training

December 11, 2023

Accelerating the training and inference processes of deep learning models is crucial for unleashing their true potential and NVIDIA GPUs have emerged as a game- Read more…

Synopsys Eats Ansys: Does HPC Get Indigestion?

February 8, 2024

Recently, it was announced that Synopsys is buying HPC tool developer Ansys. Started in Pittsburgh, Pa., in 1970 as Swanson Analysis Systems, Inc. (SASI) by John Swanson (and eventually renamed), Ansys serves the CAE (Computer Aided Engineering)/multiphysics engineering simulation market. Read more…

Some Reasons Why Aurora Didn’t Take First Place in the Top500 List

May 15, 2024

The makers of the Aurora supercomputer, which is housed at the Argonne National Laboratory, gave some reasons why the system didn't make the top spot on the Top Read more…

Leading Solution Providers

Contributors

AMD MI3000A

How AMD May Get Across the CUDA Moat

October 5, 2023

When discussing GenAI, the term "GPU" almost always enters the conversation and the topic often moves toward performance and access. Interestingly, the word "GPU" is assumed to mean "Nvidia" products. (As an aside, the popular Nvidia hardware used in GenAI are not technically... Read more…

Intel’s Next-gen Falcon Shores Coming Out in Late 2025 

April 30, 2024

It's a long wait for customers hanging on for Intel's next-generation GPU, Falcon Shores, which will be released in late 2025.  "Then we have a rich, a very Read more…

Google Announces Sixth-generation AI Chip, a TPU Called Trillium

May 17, 2024

On Tuesday May 14th, Google announced its sixth-generation TPU (tensor processing unit) called Trillium.  The chip, essentially a TPU v6, is the company's l Read more…

The NASA Black Hole Plunge

May 7, 2024

We have all thought about it. No one has done it, but now, thanks to HPC, we see what it looks like. Hold on to your feet because NASA has released videos of wh Read more…

Nvidia Shipped 3.76 Million Data-center GPUs in 2023, According to Study

June 10, 2024

Nvidia had an explosive 2023 in data-center GPU shipments, which totaled roughly 3.76 million units, according to a study conducted by semiconductor analyst fir Read more…

Q&A with Nvidia’s Chief of DGX Systems on the DGX-GB200 Rack-scale System

March 27, 2024

Pictures of Nvidia's new flagship mega-server, the DGX GB200, on the GTC show floor got favorable reactions on social media for the sheer amount of computing po Read more…

GenAI Having Major Impact on Data Culture, Survey Says

February 21, 2024

While 2023 was the year of GenAI, the adoption rates for GenAI did not match expectations. Most organizations are continuing to invest in GenAI but are yet to Read more…

AMD Clears Up Messy GPU Roadmap, Upgrades Chips Annually

June 3, 2024

In the world of AI, there's a desperate search for an alternative to Nvidia's GPUs, and AMD is stepping up to the plate. AMD detailed its updated GPU roadmap, w Read more…

  • arrow
  • Click Here for More Headlines
  • arrow
HPCwire