I just finished up with the Intel Storage Builder Summit. It was six hours of monologues, slides, and panel discussions by the heavyweights in computer storage. Based upon the event title, I expected it to be a snooze fest; but I went along because I find storage interesting and because I had nothing else to do on Saturday afternoon. Honestly, it was excellent.
The organizer of this event was Shawna Meyer-Ravelli, a manager in Intel's Storage Builders Ecosystem. From a discussion with her I learned that today's panelists and speakers were chosen for not only their involvement in hot topics in the storage ecosphere, but for their personalities and ability to create a lively discussion. It paid off.
The topics generally stuck with the theme of "Where is storage now, where is business driving it, and how can we get there?". What I took away from this summit is multi-facited: The storage scientists and engineers will continue to increase speed, remove latency, and continue to refine the storage platform into a self-adjusting appliance. However, the vast untapped potential for speed improvements requires the optimization process to climb the stack;
storage engineers see vast potential for cache optimization if the applications engineers (and to a point, end-users) would communicate business-grade metadata with them: What priority is this file? How likely am I to read it again? How likely is it to be overwritten soon? How long is this file likely to live? These are all questions that the application, programmers, and users could answer and it would have an enormous positive impact in how effective caching schemes could operate. The return on investment would be very high. But the dark clouds that overshadow this effort are the facts that the storage engineers don't speak the same language as the end users and the end users don't place the same value upon taking the time to provide their storage tier with metadata. These are cultural issues and will require just the right person who speaks both languages and has miles of patience.
Between sessions, I had an opportunity to have a conversation with Michael Mesnier and James Myers, both extremely intelligent gentlemen and down to earth guys. I could really get into the work Michael is doing with Adaptive Cache Optimization. It involves a lot of digging and I like to dig.
The speakers included...
* Jonathan H. Donaldson - VP Data Center Group, Intel
* Russ Fellows - Senior Partner & Analyist, Evaluator Group
* James Myers - Director NVM Solutions Architecture, Intel
* Michael Mesnier - Principal Engineer, Intel Labs
* Dr. Murali Rajagopal - Storage Architect, VMware
Two separate panel discussions included members from above, plus...
* Colin Gallagher - Senior Director of VxRailProduct Marketing, EMC
* Jesse St.Laurent - VP of Product Strategy, Simplivity
* Rawlinson Rivera - Principal Architect, VMware
* Kiran Sreenivasamurthy - Director, Product Management and Marketing, Maxta
* Spencer Shepler - Architect, Microsoft
* Jay Cuthrell - Director Converged Platforms and Solutions, EMC
* Howard Marks - Founder and Chief Scientist, Deep Storage
* Mark May - Technologist, Storage Geek, and coworker
The organizer of this event was Shawna Meyer-Ravelli, a manager in Intel's Storage Builders Ecosystem. From a discussion with her I learned that today's panelists and speakers were chosen for not only their involvement in hot topics in the storage ecosphere, but for their personalities and ability to create a lively discussion. It paid off.
The topics generally stuck with the theme of "Where is storage now, where is business driving it, and how can we get there?". What I took away from this summit is multi-facited: The storage scientists and engineers will continue to increase speed, remove latency, and continue to refine the storage platform into a self-adjusting appliance. However, the vast untapped potential for speed improvements requires the optimization process to climb the stack;
storage engineers see vast potential for cache optimization if the applications engineers (and to a point, end-users) would communicate business-grade metadata with them: What priority is this file? How likely am I to read it again? How likely is it to be overwritten soon? How long is this file likely to live? These are all questions that the application, programmers, and users could answer and it would have an enormous positive impact in how effective caching schemes could operate. The return on investment would be very high. But the dark clouds that overshadow this effort are the facts that the storage engineers don't speak the same language as the end users and the end users don't place the same value upon taking the time to provide their storage tier with metadata. These are cultural issues and will require just the right person who speaks both languages and has miles of patience.
Between sessions, I had an opportunity to have a conversation with Michael Mesnier and James Myers, both extremely intelligent gentlemen and down to earth guys. I could really get into the work Michael is doing with Adaptive Cache Optimization. It involves a lot of digging and I like to dig.
The speakers included...
* Jonathan H. Donaldson - VP Data Center Group, Intel
* Russ Fellows - Senior Partner & Analyist, Evaluator Group
* James Myers - Director NVM Solutions Architecture, Intel
* Michael Mesnier - Principal Engineer, Intel Labs
* Dr. Murali Rajagopal - Storage Architect, VMware
Two separate panel discussions included members from above, plus...
* Colin Gallagher - Senior Director of VxRailProduct Marketing, EMC
* Jesse St.Laurent - VP of Product Strategy, Simplivity
* Rawlinson Rivera - Principal Architect, VMware
* Kiran Sreenivasamurthy - Director, Product Management and Marketing, Maxta
* Spencer Shepler - Architect, Microsoft
* Jay Cuthrell - Director Converged Platforms and Solutions, EMC
* Howard Marks - Founder and Chief Scientist, Deep Storage
* Mark May - Technologist, Storage Geek, and coworker