|
|
Recent Advances in Flash Solutions
Sangyeun Cho
Vice President at Samsung (Memory Solutions Lab)
In this keynote talk, I will deep dive into two topics related with now
common flash memory solutions: Multi-streaming and In-Storage Computing
(ISC). Multi-streaming allows upper-level software to see
"streams" - internal parallelism within a solid-state drive (SSD) - in
a disciplined way. Proper use of multiple streams segregates data
having different properties. Our study using popular NoSQL DB
engines shows that significant gains can be achieved in terms of
steady-state SSD performance and device lifetime. ISC refers to
new capabilities of running user-written codes within the SSD, helped
by a software framework (and hardware IPs). By exploiting ample
bandwidth of flash chips and processing cycles within the SSD, one can
significantly improve the system throughput by balancing work between
compute resources and/or shortening data trips. I will discuss
results of several case studies using database and data analytics
workloads.
Bio: Sangyeun Cho
received the BS degree in computer engineering from Seoul National
University in 1994 and the PhD degree in computer science from the
University of Minnesota in 2002. In 1999, he joined the System LSI
Division of Samsung Electronics Co. and contributed to the development
of Samsung's flagship embedded processor core family
CalmRISC(TM). He was a lead architect of CalmRISC-32, a 32-bit
microprocessor core, and designed its memory hierarchy including
caches, DMA, and stream buffers. In 2004, he joined the faculty
of the Computer Science Department at the University of Pittsburgh and
was promoted to the rank of tenured associate professor in 2010.
He has recently joined Samsung’s Memory Division to lead
systems-related research efforts, where he is a VP for advanced
solutions research and development. His research interests are in
the area of computer architecture and systems with particular focus on
performance, power and reliability of memory and storage hierarchy
design for next-generation data centers.
|
|
