Advances in Flash Memory Technology & System Architecture to Achieve Savings in Data Center Power and TCO Dr. John R. Busch Vice President and Senior Fellow October 18, 2013 1
Forward-Looking Statements During our meeting today we may make forward-looking statements. Any statement that refers to expectations, projections or other characterizations of future events or circumstances is a forward-looking statement, including those relating to market position, market growth, product sales, industry trends, supply chain, future memory technology, production capacity, production costs, technology transitions and future products. This presentation contains information from third parties, which reflect their projections as of the date of issuance. Actual results may differ materially from those expressed in these forward-looking statements due to factors detailed under the caption Risk Factors and elsewhere in the documents we file from time to time with the SEC, including our annual and quarterly reports. We undertake no obligation to update these forward-looking statements, which speak only as of the date hereof. 2
Overview About SanDisk Flash Trends Flash Optimized Data Center Solutions Conclusions 3
About SanDisk 4
A Global Leader in Flash Memory Storage Solutions Rankings Trailing 4 Qtr. Financials Global Operations Technology $5.6B Revenue Fabs World Class NAND Capacity $4.4B Net Cash 5,000 + 19nm Leading Process Node $0.7B R&D Investment Employees 1Ynm Shipping Qualified at 6 of the Top 7 Server & Storage OEMs Enterprise SSDs and Storage Software SanDisk Client SSD Design Wins at 11 Leading PC OEMs The Leading Retail Brand in Key Markets The Leading Retail Brand in Key Markets #1 Global Retail Revenue Share All leading smartphone & tablet manufacturers use SanDisk Close to Half of Industry Bit Output Together with manufacturing partner Toshiba Technology Leadership 4,900 + Patents 1991 2013 Financials as of Q2, 13. Net Cash = [Cash + cash equivalents + short-term & long-term marketable securities] less [debt at maturity value] as of the end of Q2, 13. Headcount & patents as of Aug., 13. NPD Estimate, Jan., 13. Estimates of the memory card & USB markets from NPD (Jan. 13) and GfK Retail and Technology, Oct., 12. Gartner: NAND Flash Supply & Demand, WW 1Q 12-4Q 14, 2Q 13. Update Jun., 13. 5
Complete Top to Bottom Integration Performance Scalability System Utilization Endurance Cost Life Cycle Full Stack Enables Segment Optimized Solutions NAND TECH NAND DIE WAFER SCALE MFG CONTROLLER SSD SOFTWARE 6
Flash Trends 7
Toshiba SanDisk Flash Partnership Bits per cell ~4F 2 cell 3D NAND & BiCS BiCS = Bit Cost Scalable Physical scaling: 210nm 160 130 90 70 56 43 32 24 19nm Logical scaling: X1 (SLC) X2 (MLC) X3 Toshiba-SanDisk 8
2D-NAND Scaling Considerations 19 nm 19/26nm 1Y 19/19.5nm 1Z Note: Diagram not to scale 1Y technology node is 19/19.5 nm a substantial cell area reduction 1Y incorporates several new process modules to improve performance and reliability 1Z will leverage 1Y innovations to scale both X, Y dimensions substantial cell area reduction 9
3D NAND--Alternatives to Planar NAND 3D NAND Technology BiCS 3D charge trap structure Doesn t require EUV Regular optical tools Bridge technology to 3D ReRAM Bit Lines Source Lines Select Gates Word Lines Back Gate Y-cut: WL X-cut: BL Note: Diagram not to scale 10
3D Resitor RAM : will follow 3D-NAND BiCS 3D ReRAM can scale to below 10nm node providing cost reduction beyond 2020 11
Spectrum of Memory Technologies 12
Capacity (bits) Progression of Memory Technologies 1T Positioning/Prospects of Memories ReRAM BiCS HDD 100G Data Storage FG-NAND 10G 1G Working Memory DRAM 100M 10M SRAM S T T MRAM PCM NOR Code Storage ReRAM and BiCS are the two most promising post-2d NAND candidates 1M Write/Program Cycle Time (s) 1E-9 1E-8 1E-7 1E-6 1E-5 1E-4 1E-3 1E-2 Source: SanDisk, presented at IMEC 2011 13
Flash Optimized Data Center Solutions 14
Flash is Enabling New Applications, Growing Fast TAM ($B) $50 $40 $38.7 $40.2 $38.3 $32.9 $30 $24.4 $26.2 $28.5 $23.6 $20 $13.2 $10 $0 DRAM HDD NAND 2008 2012 2016E Source: Gartner 15
Total Cost of Ownership ($) Flash Drives Savings in the Enterprise SLC Lower Cost SSD MLC Caching Server/SAN Consolidation Hot, warm, and cold data High performance HDD Flash DIMM In-Memory Computing 3D Cold Storage Capacity HDD Time 16
Software Unlocks Flash Potential in the Enterprise Data Center New SAN Architecture Big Data Analytics In-Memory Databases FLASH + SOFTWARE Virtualization & Cloud Computing Cold Storage Server-side Caching Flash as replacement for 15k RPM HDD 17
Flash-Optimized Applications Flash-optimized applications: Exploit the high capacity, low latency, persistence and high throughput of flash memory Have extensive parallelism to enable many concurrent flash accesses for high throughput Use DRAM as a cache Get in-dram performance at in-flash capacity and cost, enabling server consolidation Many applications realize limited benefits from flash without system level optimization SanDisk Flash Data Fabric (FDF) is a substrate for flash-optimized applications Caching, key-value stores, databases, message queues, custom apps Leverages flash for high performance, high availability Enables low TCO through high server consolidation Executes on bare metal or virtualized 18
SanDisk Flash Data Fabric (FDF) Enables Direct Flash Access for In-Memory performance Applications Middleware Operating System Driver Firmware Flash Data Fabric (FDF) Optimizes to fully exploit flash and multi-core Encapsulates optimizations for use by any application Open and standard initiative Hardware 19
Flash Data Fabric features Provides an object API: create, replace, update, delete, indexes, range queries, transactions, snapshot Provides multiple namespaces via containers Maps object keys to flash locations Intelligent granular DRAM caching Heavily optimized access paths for high performance Optimized threading to maximize concurrency and minimize response time Configurable flash management algorithms to optimize different workloads Integrates with flash devices Minimizes write amplification, fast persistence, vectored operations, Executes in user space and is linked in as a dynamically loaded library 20
Replication Config Fault Detection Failure Handling Flash Data Fabric Architecture Databases/Data Stores Data Grid and Object Stores Application Layer Message Queue Session Store Custom Apps Container Mgmt Naming, create, open, delete Object Mgmt Naming, create, search, update, delete, FDF Protocol Layer Cluster Mgmt Naming, configure Cluster Services Local DRAM Caching Flash Manager Replication Elasticity Module Threading Module Messaging Subsystem Connect, send, receive Transport Layer 21
No SQL DataBase Example : Cassandra Cassandra is an open source distributed key-value store Key features: support for large scale synchronous and asynchronous replication, including across data centers automatic fault-tolerance and scaling tunable consistency (from writes never fail to block for all replicas to be readable ) efficient support for large rows (1000 s of columns) CQL (SQL-like) query language supports multiple indices FDF-Cassandra prototype based on Cassandra 2.1.4 22
Cassandra Performance 95/5 workload Stock Cassandra FDF Cassandra Hard Drives 64GB Data (fits in memory) 256GB Data (data set in flash) 1.2k tps 100% HDD utilization 1 of 16 cores utilization 40K tps 12 of 24 cores utilization 25K tps 90% flash utilization 18 of 24 cores utilization Intel Westmere server with 2 x 2.9GHz sockets, 24 cores, 96G DRAM SSD: 8 x 200G SSD with software RAID 0 YCSB Benchmark set-up Remote client with 10G network connection 1K fixed object, uniform distribution with configurable read/write mix (eg: 95% read, 5 % update) 48GB FDF DRAM cache N/A 124K tps 18 of 24 cores utilization 95K tps 90% flash utilization 19 of 24 cores utilization 23
TCO : Cassandra Requirement : 80k TPS and 1 TByte data set TCO - Log Scale $1,000,000 $378,216 $100,000 $55,620 3 Year OpEx 3 Year CapEx $ 14,124 $10,000 Source: Based on internal testing Stock Cassandra on HDD stock Cassandra in DRAM FDF-Cassandra and Flash 24
Transac ons per second (Thousand) FDF + Flash Accelerates Database Performance at a Dramatically Lower TCO 100 90 Yahoo! Cloud Serving Benchmark 95% Read 5% Write 95 Servers needed for 3.1TB dataset 80 70 80 60 50 40 30 20 10 0 2.4 HDD DRAM SanDisk Capacity Performance Measurements on identical commodity x86 servers; scaled with modeling from 1.0 to 3.1TB Servers are running Apache Cassandra Database Management System HDD DRAM SanDisk No. of servers 34 6 1 Power (kw) 12.7 2.8 0.4 $ per transaction $8.44 $2.49 $0.51 25
10 110 210 310 410 510 610 710 810 910 1010 1110 1210 1310 1410 1510 1610 1710 1810 1910 2010 2110 2210 2310 2410 2510 2610 2710 2810 2910 3010 3110 3210 3310 3410 3510 In Memory Data Grid Example: CouchBase vs FDF-Memcached 500000 450000 400000 350000 300000 TPS 250000 200000 Couchbase FDF-Memcached 150000 100000 50000 0 26
In Memory Database Example: FDF-Redis Performance KTPS 140 132 Bare Metal 120 116 114 100 80 60 84 101 93 70 99 93 89 Stock Redis (in memory) FDF-Redis (out of memory) 40 20 Source: Based on internal testing 0 String Hash List Set Sorted Set FDF-Redis throughput with data set in Flash matches Stock -Redis throughput with data set in DRAM 27
TCO : Stock Redis vs FDF-Redis Requirement : 80k TPS and 1 TByte data set 3 Year TCO Bare Metal 3 Year TCO AWS $250,000 350,000 $200,000 300,000 $150,000 $100,000 $50,000 3 Year OpEx 3 Year CapEx 250,000 200,000 150,000 100,000 50,000 TCO $- Stock Redis with DRAM 96GB Servers FDF-Redis with Flash 0 Stock-Redis AWS in DRAM FDF-Redis AWS with SSD Source: Based on internal testing 28
Cloud Storage Example: FDF-Swift vs Stock-Swift Performance TPS in Ks 250 FDFSwift In-memory FDFSwift in-flash StockSwift in-memory StockSwift in-flash 200 222 150 149 134 100 93 80 50 0 40 17.261 19.672 19.493 5.8 9.852 10.431 9.242 2.628 2.635 2.632 1 8 16 32 64 Clients Source: Based on internal testing, October 2013 29
Stock-Swift in Flash Client # TPS in Ks Stock Object server CPU IO Utilization FDF-Swift in Flash Client Number 1 2 1000% 15% 8 9 2400% 45% 16 10 2400% 45% 32 9 2500% 45% TPS in Ks FDF Object server CPU DRAM Miss rate IO Utilization 1 2.6 95% 30% 5% 16 40 430% 58% 48% 32 80 1048% 58% 75% 64 134 2094% 58% 92% Source: Based on internal testing, October 2013 30
Conclusions 31
Conclusions Flash technology trends will reduce TCO below all competing storage technologies Many applications realize limited benefits from flash without optimization Flash optimization of applications can yield near in-dram performance with balanced server with datasets residing in flash Broad new set of flash use cases covering entire data center: hot to warm to cold data, 32
Thank you! 2013 SanDisk Corporation. All rights reserved. SanDisk, SanDisk Ultra, SanDisk Extreme and SanDisk Extreme Pro are trademarks of SanDisk Corporation, registered in the United States and other countries. Lightning is a U.S. registered trademark of SanDisk Enterprise IP LLC. inand Extreme is a trademark of SanDisk Corporation. ULLtraDIMM is a trademark of SanDisk Enterprise IP LLC. The SD and the SDHC mark and logo are trademarks of SD-3C, LLC. Other brand names mentioned herein are for identification purposes only and may be the trademarks of their respective holder(s). 1GB=1,000,000,000 bytes. Actual user capacity less. 33