Matthew O'Keefe

Designing Storage Systems for Massive Scale

Blog Post created by Matthew O'Keefe on May 12, 2015

The 31st IEEE International Conference on Massive Storage Systems

and Technology (MSST) will be held at Santa Clara University from May 30th to June 5, 2015.

Details can be found here.


I've had the privilege of serving as the Program Chair and working with an amazing

organizing committee for the last several years, after having been a long-time participant.

HDS has been a long-time sponsor of the conference and a strong technical presence

recently in the areas of optical and FLASH. Participants from HDS this year include Ken

Wood and Michael Hay.


 

Since the conference was founded by the leading national laboratories, MSST has been a venue for

massive-scale storage architects, operators, researchers, and vendors to discuss building and

securing the world's largest storage systems for high-performance computing, web-scale systems,

and enterprises. The conference excels in creating conversations between participants designing

massive-scale storage systems via a format that includes tutorials, invited talks and panels that includes

leading practitioners and vendors, and cutting-edge research papers on massive scale systems and

related areas like FLASH.


There is a lot to talk about in this year’s program. Here are a few highlights:

 

* The tutorials track includes sessions on OpenStack, Akanda networking, Pivotal, Swift, Ceph,

and Lustre. Get some clean Linux VMs installed on your laptop and be ready to get hands-on if

you like, these tutorials will dive into a lot of detail on how to plan, deploy, and maintain these

kinds of systems.


 

* The invited program track includes some great talks including:


  — Ian Corner from CSIRO Australia will discuss designing storage systems to optimize for specific

  workflows

 

  — Storage innovators at very large data centers including NASA, BP,  Optum/UHG,

  CERN, ECMWF, and TACC will talk about their current systems and challenges, and how

  they expect their use of tape, disk, and FLASH to evolve in the next few years


 

  — Two panels on FLASH: one on integrated FLASH-based systems for large problems in high performance

  data computing, and another that focuses on how to best leverage FLASH in very large enterprise

  and web systems


 

  — Two other panels will discuss scalability in object storage and software-defined storage systems.

  Participating companies include Amazon, HDS, Nutanix, Scality, HGST/Amplidata, Seagate,

  EMC and others


 

  — Other highlights includes an overview of machine learning techniques and related storage requirements;

  a briefing on HP’s “The Machine”; and a talk from Intel on their vision for exascale storage.


 

* The research papers track this year is outstanding, including papers on building realistic web benchmarks,

file-system snapshot techniques for database and VM workloads (Netapp), numerous papers on FLASH

and Storage Class Memory innovations and programming, Baidu’s innovative new key-value store, storage

optimizations for virtualization, and more. If you want to spend a few days getting a intense update on

cutting-edge technologies for very large systems, the 31st MSST is one way to do it.


 

You can register for the conference at the MSST conference web site. Early registration ends on May 24th.

I look forward to seeing you at the conference.

Outcomes