Intel® Intelligent Storage Acceleration Library (Intel® ISA-L)

Intel® ISA-L on 01.org

Intel ISA-L Webinar - Storage Algorithms Built for Speed

YouTube* Playlist for Intel ISA-L

Subscribe to the YouTube Channel for Intel® Software

Hi. I'm Praveen from Intel. In this video, I'm going to introduce the Intel® Intelligent Storage Acceleration Library, or Intel® ISA-L. And we will cover all the key characteristics of it.

The Intel ISA-L library provides low-level functions for data protection and integrity, compression, hashing, and encryption. Don't forget to check out the links below for more information.

Let's start with a brief background of ISA-L [sic]. It's a set of low-level functions bundled up in a library that are optimized for Intel® architecture. It's a hard-boiled, bare-knuckle assembly. And that comes with a lot of advantages for storage developers.

The ISA-L [sic] is a library of highly optimized code natively designed for software-defined storage applications. This means that the library can run on any Intel CPU [sic], from Atom to Xeon [sic], from the last generation to the next generation.

There are a lot of performance advantages to be found from these optimized libraries. They are extremely efficient and fast. There are also a few novel techniques we use to get the most out of the hardware.

Next comes the future backward-compatibility and multibinary support. This is essentially the first time you call an API where it sets a jump buffer. And from then on, you will jump into the right section of the code for the processor that you run on. It queries the CPU capabilities, finds out how good it is, and uses the right binary. All you have to do to get your application moving forward to the next processor is just compile your code and run it.

It is a bare-knuckle assembly. So all you need is a platform that can run on an assembler. We have validated this on Linux, BSD, and Windows Server* operating systems. The key aspect is to use the license. People can get this code and use it for anything—commercial, open source, and more.

Intel ISA-L breaks down into five key pieces. The key five pieces are data integrity, security, encryption, compression, and data protection. The first two areas of Intel ISA-L implemented algorithms are the data protection and the data integrity areas. These act as a blocking and tackling and the passing and kicking of the storage. These are your CRC and RAID algorithms. They are the things that help build storage applications.

The second piece is the cryptographic caching. This is a translation from a random chunk of data to a fingerprint on the data. Cryptographic hashing means you cannot work backwards from hash to the data, but also [INAUDIBLE] to the interesting properties for collision resistance. In other words, if you have a large number of objects that you have to hash, you might have two hashes that have the same value. In other videos in this series, we will dive deeper into cryptographic hashing.

And next, there is encryption. The key point is that they are BSD-licensed and people can use it in commercial software.

Finally, there are the compression algorithms, or what we call igzip. We will cover this more in depth in this Intel ISA-L playlist.

Thanks for watching. Remember to check out the rest of this Intel ISA-L video playlist, where I'll be walking through some examples on compression, [INAUDIBLE] hashing, and erasure coding implementations of Intel ISA-L. Follow the links to find out more information about Intel ISA-L. And don't forget to like this video and subscribe.