Inside Solid State Drives (SSDs)

495 Pages·2018·17.41 MB·English

by  Rino Micheloni

Springer Series in Advanced Microelectronics 37
Rino Micheloni · Alessia Marelli
Editors
Kam Eshghi
Inside Solid
State Drives
(SSDs)
Second Edition
Springer Series in Advanced Microelectronics
Volume 37
Series editors
Kukjin Chun, Seoul, Korea, Republic of (South Korea)
Kiyoo Itoh, Tokyo, Japan
Thomas H. Lee, Stanford, CA, USA
Rino Micheloni, Vimercate (MB), Italy
Takayasu Sakurai, Tokyo, Japan
Willy M. C. Sansen, Leuven, Belgium
Doris Schmitt-Landsiedel, München, Germany
TheSpringerSeriesinAdvancedMicroelectronicsprovidessystematicinformation
on all the topics relevant for the design, processing, and manufacturing of
microelectronic devices. The books, each prepared by leading researchers or
engineers in their fields, cover the basic and advanced aspects of topics such as
wafer processing, materials, device design, device technologies, circuit design,
VLSI implementation, and subsystem technology. The series forms a bridge
between physics and engineering and the volumes will appeal to practicing
engineers as well as research scientists.
More information about this series at http://www.springer.com/series/4076
⋅
Rino Micheloni Alessia Marelli
Kam Eshghi
Editors
Inside Solid State Drives
(SSDs)
Second Edition
123
Editors
RinoMicheloni KamEshghi
Microsemi Corporation Lightbits Labs
Vimercate, MB SanJose,CA
Italy USA
Alessia Marelli
Microsemi Corporation
Vimercate, MB
Italy
ISSN 1437-0387 ISSN 2197-6643 (electronic)
SpringerSeries inAdvancedMicroelectronics
ISBN978-981-13-0598-6 ISBN978-981-13-0599-3 (eBook)
https://doi.org/10.1007/978-981-13-0599-3
LibraryofCongressControlNumber:2018942187
1stedition:©SpringerScience+BusinessMediaDordrecht2013
2ndedition:©SpringerNatureSingaporePteLtd.2018
Thisworkissubjecttocopyright.AllrightsarereservedbythePublisher,whetherthewholeorpart
of the material is concerned, specifically the rights of translation, reprinting, reuse of illustrations,
recitation, broadcasting, reproduction on microfilms or in any other physical way, and transmission
orinformationstorageandretrieval,electronicadaptation,computersoftware,orbysimilarordissimilar
methodologynowknownorhereafterdeveloped.
The use of general descriptive names, registered names, trademarks, service marks, etc. in this
publicationdoesnotimply,evenintheabsenceofaspecificstatement,thatsuchnamesareexemptfrom
therelevantprotectivelawsandregulationsandthereforefreeforgeneraluse.
The publisher, the authors and the editors are safe to assume that the advice and information in this
book are believed to be true and accurate at the date of publication. Neither the publisher nor the
authorsortheeditorsgiveawarranty,expressorimplied,withrespecttothematerialcontainedhereinor
for any errors or omissions that may have been made. The publisher remains neutral with regard to
jurisdictionalclaimsinpublishedmapsandinstitutionalaffiliations.
Printedonacid-freepaper
ThisSpringerimprintispublishedbytheregisteredcompanySpringerNatureSingaporePteLtd.
Theregisteredcompanyaddressis:152BeachRoad,#21-01/04GatewayEast,Singapore189721,
Singapore
To my wife Sabrina, and my daughters Laura
and Greta
Rino Micheloni
To my husband Michele and my daughter
Elena for their unconditional love
Alessia Marelli
To my wife Nazila, and my daughters Elika
and Vionna, who brighten up my life with
their love
Kam Eshghi
Foreword
Error Correcting Coding for Solid State Disk Data Storage
Wireless communication had existed for half a century when Information Theory
was expounded by Claude Shannon in the Bell System Technical Journal in 1948.
Error correcting coding followed in primitive formulations which brought early
digitalcommunicationsystemsonlyashortwaytowardtheShannoncapacitylimit.
Various generations of algebraic codes: Hamming, BCH and Reed Solomon made
gradualprogress.Withtheadventofdigitalsatellitetransmissionandsoft-decision
decoding of convolutional codes, the gap between uncoded performance and the
Shannon limit was cut in half. Similar technology was used in second and third
generation(2Gand3G)mobilephonevoicemodems.Finallyturbocodesandlow
densityparitycheck(LDPC)codes,whicharrivedabouttwodecadesago,gradually
were shown to greatly decrease the distance to the capacity limit. These tech-
nologieshaveenteredpredominantusefordatatransmissionin3Gand4Gmobile
modems.
High density data storage technology has followed a similar trajectory though
with a more contracted time span. BCH and Reed Solomon codes were the norm
until recently for hard disk drives (HDD). Recently though LDPC has taken root
here too with major improvements in data density and reading and writing con-
trollerspeeds.Withtheadventofthe“smartphones”andtablets,solidstatedrives
(SSD)becameevermoreimportantfortheirlowlatencyandlowpoweroperation.
For this use LDPC is becoming the norm as well. This book which covers all
aspects of SSD technology also provides coverage of the important topic of ECC.
La Jolla, CA, USA Andrew Viterbi
President
Viterbi Group, LLC
vii
Preface to the Second Edition
WestartedwritingthefirsteditionofInsideSolidStateDrives(SSDs)backin2011,
andthebookwasfirstpublishedin2013.Atthattime,SSDswereconsideredasthe
“new” technology in the storage space, but not really a “shining star” as they are
seen today.
Overthepastfewyears,wehavecollectedalotoffeedbackandquestionsabout
our book. Moreover, both SSD and Flash technologies have significantly changed
alongtheway.Therefore,wethoughtitwastherighttimetorefreshthecontentof
“Inside Solid State Drives.”
As editors, we have pushed all co-authors to refresh each chapter (Thank You
ALL!), in terms of both the content and the bibliography.
But this second edition is much more than that.
Asmentioned,SSDtechnologieshavesignificantlychangedinthelastfiveyears
and we realized that there was the need to add three completely new chapters:
Chaps. 5, 7, and 9.
In 2013, Flash manufacturers were still fighting against the challenges of
shrinking the size of planar memory cells to keep up with the expectations of the
market in terms of $/bit. Now, Flash technology is 3D (i.e., vertically integrated)
and there is a new dimension to consider: the number of memory layers (100+ in
the near future). 3D NAND Flash appeared in the market at the end of 2015, but
there is still a plethora of alternatives around, based on different architectures and
memory technologies (floating gate and charge trap). Chapter 5 covers 3D Flash
array architectures with a lot of bird’s-eye views, to help the reader understand
better the new challenges that technologists and developers have to face.
In all SSDs, a Flash microcontroller sits between one or multiple hosts (i.e.,
CPUs) and NAND Flash memories, and on each side, there are a lot of challenges
that designers need to overcome. Moreover, a single controller can have multiple
cores, with all the complexity associated with developing a multi-threaded firm-
ware. Chapter 7 is about how to make simulations of such a complex system, by
providing insights into design trade-off and simulation strategies. As usual, simu-
lation speed andprecision donotgohandinhand,soitisimportanttounderstand
ix
x PrefacetotheSecondEdition
when to simulate what. Of course, being able to simulate SSD’s performances is
necessary to meet time-to-market, as well as price and quality targets.
Nowadays, SSDs are electronic systems much more complex than in the past,
especially because they have to manage a lot of 3D memories, by using several
algorithms (wear leveling, Error Correction Code, soft decoding, randomization,
read retry, etc.) at a very high speed (especially with PCIe/NVMe drives).
Chapter 9 is exactly designed to offer a comprehensive overview of the most
recent Flash management techniques (aka Flash Signal Processing). We are sure
that technologists, engineers, and scientistwillappreciate theunbelievablelevel of
know-how required by the management of electrons and holes inside nonvolatile
memory cells.
We really placed our best effort in updating this book. Enjoy the reading!
Vimercate, Italy Rino Micheloni
Vimercate, Italy Alessia Marelli
San Jose, USA Kam Eshghi
Preface to the First Edition
Solid State Drives (SSDs) are gaining momentum in enterprise and client appli-
cations, replacing Hard Disk Drives (HDDs) by offering higher performance and
lowerpower.Intheenterprise,developersofdatacenterserverandstoragesystems
haveseenCPUperformancegrowingexponentiallyforthepasttwodecades,while
HDD performance has improved linearly for the same period. Additionally,
multi-core CPU designs and virtualization have increased randomness of storage
I/Os. These trends have shifted performance bottlenecks to enterprise storage sys-
tems. Business critical applications such as online transaction processing, financial
data processing and database mining are increasingly limited by storage
performance.
Inclientapplications,smallmobileplatformsareleavinglittleroomforbatteries
while demanding long life out of them. Therefore, reducing both idle and active
powerconsumptionhasbecomecritical.Additionally,clientstoragesystemsarein
need of significant performance improvement as well as supporting small robust
formfactors.Ultimately, clientsystemsareoptimizingforbestperformance/power
ratio as well as performance/cost ratio.
SSDs promise to address both enterprise and client storage requirements by
drastically improving performance while at the same time reducing power.
Inside Solid State Drives walks thereader through all the main topics related to
SSDs.
A Solid State Drive is a very complex system: Chapter 1 contains an overview
of the main blocks, including hardware and software.
Chapters 1 and 2 cover different SSD implementations with host interfaces
ranging from SAS/SATA to PCI Express (PCIe). SAS/SATA offer compatibility
with legacy storage infrastructure. However, for many applications, NAND Flash
read and write speeds are exceeding the capabilities of these legacy interconnects.
PCIeSSDsovercomethisbottleneckanddeliverunparalleledperformancewhile,at
the same time, reducing latency, power and cost by eliminating the traditional
storage infrastructure and attaching directly to a platform’s PCIe I/O interconnect.
xi