View
18
Download
4
Category
Tags:
Preview:
Citation preview
Accelerating Applications UsingFlash and Non-Volatile Memory
Saeed RajaDirector Product Marketing Management
SanDisk Corporation
Flash Memory Summit 2015Santa Clara, CA 1
Forward-Looking Statements
During our meeting today we will 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 industry trends, future products and technologies and product performance capabilities and availability.
Actual results may differ materially from those expressed in these forward-looking statements due to a number of risks and uncertainties, including the 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.
Flash is a Major Disruptor in Datacenters…Non-Volatile Memory has Made Big News Recently!!!
NAND Flash Major Trends:
− Capacity: 100s of GB to 100s of TB per device
− Higher capacity, lower cost/GB− Lower write cycles− SLC->MLC->3BPC− IOPS: 100K to millions− GB/s of bandwidth
Non-Volatile Memory DDR/PCI-e attached NVDIMM / Capacitance
backed power-safe buffers + FLASH Orders of magnitude performance improvement
SAS and SATA attach SSDs
PCIe attached Fabric attached
NVDIMMsNext Phase :: Persistence Memory
Flash DevicesMemory Persistent Memory
Non Volatile Memory File System (NVMFS)
Persistent Memory Manager (PMM)
Memory Interfaces
Applications
No Changes in Applications Applications Optimized
“Flash -Aware”“Transparent”
Application Acceleration Using Flash Extended Memory
2xPerformance
Increase
4xPerformance
Increase
Native Flash Translation Layerblock allocation, mapping, recycling
ACID updates, logging/journaling, crash-recovery
NVMFSfile metadata mgmt
Kernel block layer
kernel-space
user-space
Standard Linux File Systemsfile metadata mgmt,
block allocation, mapping, recycling,ACID updates, logging/journaling,
crash-recovery
NVM Primitives
MySQL™ Database Use Case
Linux Virtual File System - Abstraction Layer
New File System
NVMFS is POSIX compliant
Leverages the functionality of underlying Flash Translation Layer (FTL)
Namespace Management
Enables Direct flash/memory access and crash recovery
NVM primitives are exposed through standard system file interface
NVMFS - Optimized for Flash and Persistent Memory
Flash Devices
Memory Persistent Memory
Every MySQL write translates to 2 writes to storage device
80% performance penalty with legacy MySQL compression
enabled
Page CPage
B
Page A
Buffer
DRAMDRAMBufferBuffer
SSD (or HDD) Database
Database Server
Page C
Page B
Page A
Page C
Page B
Page A
Page C
Page B
Page A
Application initiates updates to pages A, B, and C.
1
MySQL copies updated pages to memory buffer.
2
MySQL writes to double-write buffer on the media.
3
Once step 3 is acknowledged, MySQL writes the updates to the actual tablespace.
4
2 Writes
80
% r
ed
uct
ion
in T
PS
21
Tra
nsac
tion
Rat
e
Results and performance may vary according to configurations and systems, including drive capacity, system architecture and applications.
Legacy MySQL Challenges
MySQL with Atomic Write
Fusion ioMemory Database
Page C
Page B
Page
A
DRAMBuffer
Page C
Page B
Page
A
Application initiates updates to pages A, B, and C.
1
MySQL copies updated pages to memory buffer.
2
MySQL writes to actual tablespace, bypassing the double-write buffer step due to inherent atomicity guaranteed by the intelligent Fusion ioMemory device.
3
Database Server
Page CPage
B
Page
A
One, Single Atomic Write!One, Single
Atomic Write!
Solving the Double-Write Challenge with NVMFS
Enhanced Life Expectancy of Flash Devices:– Reduce Writes to flash by half at similar throughput
Improved performance consistency
Reduced latency, increased transaction/sec
Higher performance – Especially workloads with datasets that are bigger than DRAM
Perfect Fit for ACID-compliant MySQL
Changes have been contributed to the MySQL community.Oracle MySQL 5.7.4, Percona Server 5.5, 5.6 and MariaDB 10
Atomic Write significantly reduces latency while Increasing performance consistency
Sysbench - MariaDB 10.0.15, 4000 OLTP TXN injection/second, 99% latency, 220GB data - 10GB buffer pool
NVMFS latency range
XFS latency range
Significantly Improved Latency Consistency
Benefits of compression without performance penalty Within 10% of uncompressed
Up to 50% improvement in capacity utilization
Enhanced life expectancy of flash devices Up to 4x fewer writes to storage with
Compression and Atomic Write Tra
nsac
tion
Rat
eThe performance results discussed herein are based on internal testing and use of Fusion ioMemory products. Results and performance may vary according to configurations and systems, including drive capacity, system architecture and applications.
Significantly Improved MySQL Compression
Compression with almost no performance penalty
Changes have been contributed to the MySQL community.Oracle MySQL 5.7.4, Percona Server 5.5, 5.6 and MariaDB 10
TPC-C-Like benchmark, 1000 warehouses - 75GB Buffer pool, MariaDB 10.0.15
5x improvement
Legacy MySQL with Compression ON
MySQL /NVMFS with Compression ON
Legacy MySQL with Compression OFF
NVM Compression eliminates legacy MySQL’s compression penalty
Compression without Performance Penalty
2x
2.1x
2x2x
2x
MySQL :: Transparent Acceleration
2x Higher transactions per second using DDR attached NVDIMM
Baseline = without NVDIMMMySQL = without any modifications
MySQL :: Transparent Latency Improvements
NVMFS provides consistent low latency
Baseline = without NVDIMMMySQL = without any modifications
Cassandra :: Major Reduction in WRITES to Flash
Up to 3.2x reduction in Writes to Flash---Expected Longer Endurance
Cassandra, 2.2.0YCSB BenchmarkFlash Writes = GB
Lower is Better2x
3x
2x
Baseline = without NVDIMMCassandra = without any modifications
Customer Benefits
Changes have been contributed to the MySQL community.Oracle MySQL 5.7.4, Percona Server 5.5, 5.6 and MariaDB 10
Transparent Flash AwareInsert Heavy
Throughput
Latency
Up to 2x improvements
Up to 4x improvements
Use “Flash-as-Memory” byte-addressable architecture and interface
Expected Longer Endurance of Flash Devices
Application Level Benefits
ISV’s and Partners Inquiries Welcomed!!!
Questions?
Flash Memory Summit 2015Santa Clara, CA 15
Thank You!
@BigDataFlash #bigdataflash
ITblog.sandisk.comhttp://bigdataflash.sandisk.com
For details about NVFMS refer to:
NVMFS: A New File System Designed Specifically to Take Advantage of Nonvolatile Memory.Dhananjoy Das (Sr. Systems Architect SanDisk Corp.) in Forum D-21: Enterprise Storage Design.
Recommended