1 © 2013 ibm corporation aprovechando el valor de la información crucial en la organización...
TRANSCRIPT
1 © 2013 IBM Corporation
Aprovechando el valor de la información crucial en la organización
FlashSystem Family
Adrian RestucciaHigh End Storage Technical [email protected]
© 2013 IBM Corporation2
2009
800,000 petabytes
2020
35 zettabytesas much Data and ContentOver Coming Decade
44x Business leaders frequently make decisions based on information they don’t trust, or don’t have
1 in 3
83%of CIOs cited “Business intelligence and analytics” as part of their visionary plansto enhance competitiveness
Business leaders say they don’t have access to the information they need to do their jobs
1 in 2
of CEOs need to do a better job capturing and understanding information rapidly in order to make swift business decisions
60%Of world’s datais unstructured
80%
Information is at the Center of a New Wave of Opportunity…
… And Organizations Need Deeper Insights… And Organizations Need Deeper Insights
Smarter Computing
© 2013 IBM Corporation3
Consider Big Data within an enterprise’s information supply chain and storage infrastructure.
External Information
Sources
Transactional & CollaborativeApplications
Business Analytics Applications
Smarter Computing
IBM FlashSystems
4 © 2013 IBM Corporation4
Smarter Computing Demands Flash
“The more we have flash for consumer devices...”
“...the more we need flash for our data centers...”
5 © 2013 IBM Corporation5
Why Flash Storage…….. Timing is Perfect!
In the last 10 years…
CPU Speed: Performance increase roughly 8-10x
DRAM Speed: Performance increase roughly 7-9x
Network Speed: Performance increase of 100x
Bus Speed: Performance increased roughly 20x
Disk speed: Performance increased 1.2x
IBM FlashSystem™
© 2013 IBM Corporation6
IT Infrastructure Challenges
From 1980 to 2010, CPU performance has grown 60% per year*
…and yet, disk performance has grown ~5% per year during that same period**
From 1980 to 2010, CPU performance has grown 60% per year*
…and yet, disk performance has grown ~5% per year during that same period**
CPU performance has grown 10x in the last decade
While storage has grown capacity it has been unable to keep up in performance
Systems are now Latency & IO bound resulting in a significant performance gap
CPU performance has grown 10x in the last decade
While storage has grown capacity it has been unable to keep up in performance
Systems are now Latency & IO bound resulting in a significant performance gap
* IBM study of CPU performance ** IBM study of disk performance
Performance Gap
© 2013 IBM Corporation7
Most Costly & Volatile
Time Consuming, Very Expensive &
Risky
Wasteful, Expensive & Ineffective with Storage Latency
Issues
Expensive & Ineffective for
Storage Performance Issues
Client Responses to Performance Gap
Add More Memory
Typical Performance
Mitigation Tactics
HDD Performance Enhancement
Add CPUsTune & Modify
Application
© 2013 IBM Corporation8
All Flash is About Economics
Improve Performance
ReduceCosts
Enable NewOpportunities
Leverage the “Economies of Scale” of Flash
– Accelerate Application Performance– Gain Greater System Utilization – Lower Software & Hardware Cost– Save Power / Cooling / Floor Space– Drive Value Out of Big Data
© 2013 IBM Corporation9
No applicationOr architecture Changes
Benefits & economics outweigh disk
Reduce floor space, power & cooling
Servers, Applications and Databases are FASTER!
Understanding Application Efficiency using FlashSystem
CPU Utilization & App. Efficiency
4%Total Application Processing Time 5,200us (5.2ms)
5,000us (5ms)
200us (.2ms)
Application Processing Time
Time Waiting for I/0(Waiting for Array)
Time Processing Data(Server CPU)
Disk/Hybrid/SSD IBM FlashSystem
CPU Utilization& App. Efficiency
50%Total Application Processing Time400us (.4ms)
200us (.2ms)
200us (.2ms)
What do you do with the Extra
Time?
© 2013 IBM Corporation10
Race to Zero Latency
Tape DrivesHard Drive Disks
Solid State Disk
IBM FlashSystem
Seconds5-15 milliseconds
~1 milliseconds
200 microseconds, 1U up to 40 TB
Zero Latency
With each new gen. of storage, comes performance gains in order of magnitude
© 2013 IBM Corporation11
•Supplement your existing infrastructure.
•Assign the heaviest, most critical workload to IBM FlashSystem.
•Decrease overall response times.
•Increase efficiency/utilization across the IT stack.
•Completely eliminate storage performance issues
IBM FlashSystem – Remove I/O BottleNeck
© 2013 IBM Corporation12
•IBM FlashSystem attaches to a Server, SAN Fabric, or SVC•IBM FlashSystemappears as a block storage device
•Installation in under 60 minutes•Slide into Rack/attach FC or IB•Plug in Ethernet for GUI•Define LUNS / Allocate files•No RAID Groups/No Tuning Controls•Management Port provides real time statistics
IBM FlashSystem –Easy Deployment
© 2013 IBM Corporation13
Industry Vertical
Use Case
Where Do You Best Use Flash Today
Typical Use Cases and Verticals
Cloud-Scale Infrastructures
OLTP Databases
Virtual Infrastructures
Computational Applications
Analytical Apps/OLAP
Government
Financial
Telecom
High Performance Computing
eCommerce
© 2013 IBM Corporation14
Where Do You Best Use Flash Today
© 2013 IBM Corporation15
Exadata Example
© 2013 IBM Corporation16
ASM Preferred Read
© 2013 IBM Corporation17
HANA Example
© 2013 IBM Corporation18
HANA Example
© 2013 IBM Corporation19 © 2014 International Business Machines Corporation
CAPI – an open invitation to innovate on POWER
FPGA
Power Processor
Coherent Attached Processor Proxy (CAPP) in processor
• Unit on processor that extends coherency to an attached device
• On processor directory responds on behalf of off-chip device (Filtering snoops)
Coherency protocol tunneled over standard PCIe• Eliminates the need for special I/Os and protocol logic
– CAPI utilizes standard Posted Write and Non-posted Reads
• Reduces the complexity and bandwidth requirements of the attached device
Enables attached device to be a peer to the processor
• Simplifies programming model between application• Enables device to use same effective address as application
running in processor• Eliminates the cumbersome I/O Device Driver requirements
– Pinned memory not required
POWER8 Synergy – CAPI ( Coherence Accelerator Processor Interface)
© 2013 IBM Corporation20 © 2014 International Business Machines Corporation
CAPI Attached Flash Optimization
Issues Read/Write Commands from applications to eliminate 97% of instruction path length CAPI Flash controller Operates in User Space
Pin buffers, Translate, Map DMA, Start I/O
Application
LVM
Disk & Adapter DD
Read/WriteSyscall
strategy() iodone()
FileSystemstrategy() iodone()
Interrupt, unmap, unpin,Iodone scheduling
< 500 Instructions < 500 Instructions
ApplicationPosix Async I/O Style API
User LibraryShared Memory Work Queue
aio_read()aio_write()
20K Instructions 20K Instructions
Attach flash memory to POWER8 via CAPI coherent Attach
POWER8 Synergy – CAPI ( Coherence Accelerator Processor Interface)
21 © 2013 IBM Corporation
IBM FlashSystem Family
FlashSystem V840 FlashSystem 840
22 © 2013 IBM Corporation
IBM FlashSystem 840 Data Center Optimized
• 1.1M IOPS• 8 GB/s Bandwidth• Multiple connectivity interfaces
- 16Gb/8Gb Fibre Channel- 40Gb QDR InfiniBand- 10Gb FCoE
• Fully redundant and hot swappable architecture: - Flash modules, power supplies, batteries, interfaces, fans
• Non-disruptive maintenance and updates - Concurrent code load, highly serviceable design
• Encryption
• 2U form factor- minimal footprint for best of breed ROI• Low power 625 watts• Field upgradeable, granular capacity
- 4, 8, 12, 16, 20, 24, 32, 40, 48• Reduce installation and management time
with intuitive standardized GUI
• Low Latency 135/90 µs R/W • Purpose-built, highly parallel design• Maximize host CPU efficiency and productivity
Macro Efficiency
MicroLatency™
Enterprise Reliability
Extreme Performance
23 © 2013 IBM Corporation
IBM FlashSystem 840: Hardware View
Flash Modules (12)
RAID Controllers (2)
Battery Modules (2)
Power Supplies (2)
Fan Modules (4)
Interface Modules (4)
Management Modules (2)
Canisters (2)
Improved RAS featuresFront/Back accessible Hot-swap Flash Modules, Power Supplies, Batteries, Fans, Controllers w/ interface cards and CanistersNon-disruptive maintenance and firmware updates (concurrent code load)
Improved RAS featuresFront/Back accessible Hot-swap Flash Modules, Power Supplies, Batteries, Fans, Controllers w/ interface cards and CanistersNon-disruptive maintenance and firmware updates (concurrent code load)
© 2013 IBM Corporation24
IBM FlashSystem Midrange SSD-based Array
Source: SPC-1/E Result #AE00006 (IBM FlashSystem 820), 8/16/2013, and SPC-1 Result #A00134 (HP 3PAR StoreServ 7400 Storage System (with SSDs)), 5/23/2013. Data indicated by * is derived from SPC data, not reported directly. More footnotes coming…
Price per GB Latency (minimum) Power Space Capacity
density (GB/U)IOPS density
(IOPS/U)
75% better
45% better
30% better
75% better
97% better
67% better
Optimization Drives Data Economics
SPC-1 Comparison – FlashSystem vs. SSDs
Under Embargo Until Nov. 19, 2013
© 2013 IBM Corporation25
IBM Flash Storage Impact on Systems Economics
85% Reductionin batch
processing times
85% Reductionin batch
processing times
50% Reductionin Software Licenses
50% Reductionin Software Licenses
80% ReductionEnergyUsage
80% ReductionEnergyUsage
75% Reductionin footprint;1 Petabyte
on1 floor tile.
75% Reductionin footprint;1 Petabyte
on1 floor tile.
Better Economics Without Re-architecting Applications
100 µs LatencyNo more
bottlenecks
100 µs LatencyNo more
bottlenecks
Enterprise ReliabilityHigh Availability, 2D Flash RAID
& IBM Variable Stripe RAIDTM
Enterprise ReliabilityHigh Availability, 2D Flash RAID
& IBM Variable Stripe RAIDTM
The data below are based on average operating conditions that may or may not be representative of a particular customer’s operating environment. The use case measurements are from TMS customers using the flash technology that has been integrated into IBM’s systems
© 2013 IBM Corporation26
John the Datacenter Owner
27 © 2013 IBM Corporation
FlashSystem Family
GRACIAS!!
Adrian RestucciaHigh End Storage Technical [email protected]