1. 1
Expanding the
Possibilities of Storage
For enterprise, hyperscale, cloud
For consumers
For mobile and connected OEMs
Rob Callaghan
Ultra-Low Latency and Scalable I/O Performance
for Accelerated Database Performance
2. 2
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,
construction schedules, production starts, 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.
3. 3
Widening
Performance Gap
1990 1995 2000 2005 2010 2015
Server CPU Performance
Challenge: Legacy Storage I/O Bottleneck
Source: StorageIOblog; Sep 2009; http://storageioblog.com/data-center-io-bottlenecks-performance-issues-and-impacts/
Aligning
Performance
SSD
Performance
HDD
Performance
4. 4
I/O Challenges Affect Productivity and Revenue
“… some people have a hard time grasping it, these (solid state)
drives save a tremendous amount of money.”
“Even very small delays would result in substantial and costly
drops in revenue” [when search results pages were slowed by
1 second … +0.1 seconds costs Amazon 1% of sales]
Source: http://highscalability.com/blog/2009/7/25/latency-is-everywhere-and-it-costs-you-sales-how-to-crush-it.html
and Cade Metz, “Flash Drives Replace Disks at Amazon, Facebook, Dropbox,” Wired, June 13
Artur Bergman
Founder of Fastly
Greg Linden
Amazon
5. 5
Industry Challenge: Efficiently Scaling Databases
Data explosion accompanied by growing
demand for transactional performance
and analytics
Adding servers / CPU / memory / HDD
storage is not cost effective
Need for low-latency I/O and storage
Opportunity: Seamless, low-latency database scaling
6. 6
Industry Challenge: Datacenter Sprawl
IT environments running up against
I/O and storage bottlenecks
Low storage performance creates
inefficiencies in datacenter
Application acceleration costs are
straining IT budgets
Opportunity: Eliminate I/O bottlenecks to allow storage consolidation
7. 7
Flash Storage - Database Performance Game Changer
Flash Storage into the Data Center provides performance gains that were
difficult to reach using hard disk technology.
SanDisk’s broad range of Flash Solutions allow companies to easily tier data
using the features of Oracle,
– Providing solutions for servers and storage arrays with different performance
characteristics and cost metrics
Intelligent Software takes advantage of this faster hardware.
Database Smart Flash Cache, on Oracle Linux to accelerate I/Os for read-mostly
database and provides intelligent caching of hot data on the Oracle server.
– Using flash as a persistent extension of system memory
• Auto-Commit Memory for Oracle
• Atomic Writes for Oracle
9. 9
The World’s First Flash Storage Device
on the Memory Channel
Enterprise level endurance with MLC
10 drive writes per day
5 year warranty
Guardian Technology™
Back up power circuitry
Full data path protection
Enterprise class MTBF
Enterprise Class Reliability
DDR3 protocol
Configured as block device
(through device driver)
Memory Channel Interface
Lowest write latency (less than 7 usec)
Scalable performance by adding additional
ULLtraDIMM™ SSDs
Ultra Low Latency, High Performance
200, 400 GB
Scalable architecture
19nm flash technology (MLC)
Scalable, Cost Effective Media
10. 10
Scalable I/O Performance, Constant Latency
0
10
20
30
40
50
60
70
80
90
100
0
200
400
600
800
1000
1200
1400
1 2 4 8
Latency(us)
(K)IOPS
Number of ULLtraDIMM™ Devices
400G
800G
1600G
3200G Demonstrated
at the Open
Compute Summit
– Jan 2014
IBM 3650 2-socket server
running Red Hat version
6.3