Enterprise Data Center Storage

StorPool Presents at Storage Field Day 18

Storage field day 18 was held in Silicon Valley between 27-th of February and 1-st of March 2019. 

Our session was on the first day and it made waves. Ray Lucchesi (@RayLucchesi) said:

“Standard, enterprise-class, million dollars storage system is not able to do this. Standard software-defined versions of these sorts of systems with tiers of SSDs aren’t able to do these numbers. We are talking about a million IOPS mix read/write workload.” Here is the full video.

Many thanks to Stephen Foskett (@SFoskett) and Gestalt IT for the great organization. Everything was perfect. Huge thanks to our partners at Intel too for building the Lab for SFD18 demo.

Did you miss the live session?

View the slides from presentation here or view the videos below:

DOWNLOAD THE SLIDES FROM THE PRESENTATION

Boyan Ivanov, a CEO and a co-founder, presents a short company introduction. In brief, StorPool is not the typical Silicon Valley company. The product was built from scratch to deliver scale-out block storage software at incredible levels of performance and really low latency. Focused on primary use cases, usually, flash systems.  

Boyan Krosnov, a Chief of Product and a co-founder, with an overview of what we are doing. There is this common knowledge in storage systems that you can’t do data management, performance and scalability at the same time. Boyan explains – “If you have a system that does 2 of these – that’s exceptional, a system that does 3 is considered impossible.”
StorPool can easily deliver on the three.

Boyan Krosnov, Architecture and Demonstration. In this video Boyan is giving an architecture deep dive of StorPool followed by performance demo. 

The next video is a detailed Performance demo of StorPool and presenting how StorPool achieves 1 million IOPS mix read/write workload and even more.

We started with an interesting document recently published from Microsoft and decided to build our demo set-up based on that, with the small difference that with StorPool, we optimize the hardware underneath. What we changed is that we removed the Intel Optane DC and the RDMA and most importantly the copies are not local but 100% remote. After dissecting the set-up the demo goes deeply technical showling live records of hitting 1 mln IOPS mix 70/30 random read/writes.

For a reference: https://blogs.technet.microsoft.com/filecab/2018/10/30/windows-server-2019-and-intel-optane-dc-persistent-memory/.  “The new HCI industry record: 13.7 million IOPS with Windows Server 2019 and Intel® Optane™ DC persistent memory”

And last, but not least, Boayn Ivanov again, with presenting a few case studies.

Share this Post

Leave a Reply

Your email address will not be published. Required fields are marked *