Storage Performance Considerations for AI Data Paths

The hype of Deep Learning (DL), Machine Learning (ML) and Artificial Intelligence (AI) has reached an unprecedented frenzy. Every infrastructure vendor from servers, to networking, to storage has a word to say or play about DL/ML/AI. This prompted me to explore this hyped ecosystem from a storage perspective, notably from a storage performance requirement point-of-view.

One question on my mind

There are plenty of questions on my mind. One stood out and that is related to storage performance requirements.

Reading and learning from one storage technology vendor to another, the context of everyone’s play against their competitors seems to be  “They are archaic, they are legacy. Our architecture is built from ground up, modern, NVMe-enabled“. And there are more juxtaposing, but you get the picture – “We are better, no doubt“.

Are the data patterns and behaviours of AI different? How do they affect the storage design as the data moves through the workflow, the data paths and the lifecycle of the AI ecosystem?

Continue reading

Lift and Shift Begone!

I am excited. New technologies are bringing the data (and storage) closer to processing and compute than ever before. I believe the “Lift and Shift” way would be a thing of the past … soon.

Data is heavy

Moving data across the network is painful. Moving data across distributed networks is even more painful. To compile the recent first image of a black hole, an amount of 5PB or more had to shipped for central processing. If this was moved over a 10 Gigabit network, it would have taken weeks.

Furthermore, data has dependencies. Snapshots, clones, and other data relationships with applications and processes render data inert, weighing it down like an anchor of a ship.

When I first started in the industry more than 25 years ago, Direct Attached Storage (DAS) was the dominating storage platform. I had a bulky Sun MultiDisk Pack connected via Fast SCSI to my SPARCstation 2 (diagram below):

Then I was assigned as the implementation engineer for Hock Hua Bank (now defunct) retail banking project in their Sibu HQ in East Malaysia. It was the first Sun SPARCstorage 1000 (photo below), running a direct attached Fibre Channel 0.25 Gbps FCAL (Fibre Channel Arbitrated Loop). It was the cusp of the birth of SAN (Storage Area Network).

Photo from https://www.cca.org/dave/tech/sys5/

The proliferation of SAN over the next 2 decades pushed DAS into obscurity, until SAS (Serial Attached SCSI) came about. Added to the mix was the prominence of Cloud Storage. But on-premises storage and Cloud Storage didn’t always come together. There was always a valley between the 2, until the public clouds gained a stronger foothold in the minds of IT and businesses. Today, both on-premises storage and cloud storage are slowly cosying as one Data Singularity, thanks to vision and conceptualization of data fabrics. NetApp was an early proponent of the Data Fabric concept 4 years ago. Continue reading

WekaIO controls their performance destiny

[Preamble: I have been invited by GestaltIT as a delegate to their Tech Field Day for Storage Field Day 18 from Feb 27-Mar 1, 2019 in the Silicon Valley USA. My expenses, travel and accommodation were covered by GestaltIT, the organizer and I was not obligated to blog or promote their technologies presented at this event. The content of this blog is of my own opinions and views]

I was first introduced to WekaIO back in Storage Field Day 15. I did not blog about them back then, but I have followed their progress quite attentively throughout 2018. 2 Storage Field Days and a year later, they were back for Storage Field Day 18 with a new CTO, Andy Watson, and several performance benchmark records.

Blowout year

2018 was a blowout year for WekaIO. They have experienced over 400% growth, placed #1 in the Virtual Institute IO-500 10-node performance challenge, and also became #1 in the SPEC SFS 2014 performance and latency benchmark. (Note: This record was broken by NetApp a few days later but at a higher cost per client)

The Virtual Institute for I/O IO-500 10-node performance challenge was particularly interesting, because it pitted WekaIO against Oak Ridge National Lab (ORNL) Summit supercomputer, and WekaIO won. Details of the challenge were listed in Blocks and Files and WekaIO Matrix Filesystem became the fastest parallel file system in the world to date.

Control, control and control

I studied WekaIO’s architecture prior to this Field Day. And I spent quite a bit of time digesting and understanding their data paths, I/O paths and control paths, in particular, the diagram below:

Starting from the top right corner of the diagram, applications on the Linux client (running Weka Client software) and it presents to the Linux client as a POSIX-compliant file system. Through the network, the Linux client interacts with the WekaIO kernel-based VFS (virtual file system) driver which coordinates the Front End (grey box in upper right corner) to the Linux client. Other client-based protocols such as NFS, SMB, S3 and HDFS are also supported. The Front End then interacts with the NIC (which can be 10/100G Ethernet, Infiniband, and NVMeoF) through SR-IOV (single root IO virtualization), bypassing the Linux kernel for maximum throughput. This is with WekaIO’s own networking stack in user space. Continue reading

Microsoft desires Mellanox

My lazy Thursday morning was spurred by a posting by Stephen Foskett, Chief Organizer of Tech Field Days. “Microsoft mulls the acquisition of Mellanox

The AWS factor

A quick reaction leans towards a strange one. Microsoft of all people, buying a chip company? Does it make sense? However, leaning deeper, it starts to make some sense. And I believe the desire is spurred by Amazon Web Services announcement of their Graviton processor at AWS re:Invent last month.

AWS acquired Annapurna Labs in early 2015. From the sources, Annapurna was working on low powered, high performance networking chips for the mid-range market. The key words – lower powered, high performance, mid-range – are certainly the musical notes to the AWS opus. And that would mean the ability for AWS to control their destiny, even at the edge. Continue reading