FLASH 1st - The Storage Strategy for the Next Decade ( EMC World 2012 )

2,071 views

Published on

This session discusses the FLASH 1st storage strategy and how FLASH is the first line of defense of a modern, dynamic storage hierarchy. We further explore new FLASH deployment models like host based PCIe FLASH and all FLASH arrays. This session also discusses a business driven sizing model that enables you to quickly and precisely estimate just how much FLASH, 15K HDD and NL HDD is needed for the optimal FLASH 1st strategy.

Published in: Technology
0 Comments
3 Likes
Statistics
Notes
  • Be the first to comment

No Downloads
Views
Total views
2,071
On SlideShare
0
From Embeds
0
Number of Embeds
2
Actions
Shares
0
Downloads
111
Comments
0
Likes
3
Embeds 0
No embeds

No notes for slide

FLASH 1st - The Storage Strategy for the Next Decade ( EMC World 2012 )

  1. 1. FLASH 1st The Storage Strategy for the Next Decade Denis Vilfort Sr. Director, USD Marketing© Copyright 2012 EMC Corporation. All rights reserved. 1
  2. 2. Information Tipping Point AheadThe future will be nothing like the past 140,000 120,000 Data A 10TB data center in 110X 100,000 2001growing at 60% YoY 80,000 will be a 120 Petabyte We data center by 2021, - but 60,000 MORE DATA are EACH DECADE IT budgets remain flat Automation 40,000 Here 20,000 Budget 0 2000 2003 2006 2009 2012 2015 2018 2021© Copyright 2011 EMC Corporation. All rights reserved. 2
  3. 3. Most Data Will be Stone ColdWhile we increasingly can’t erase data, we can store it better 4,123 110 X 37.5 TB TB20112021 Hot Warm Cold Evolution of 50 TBs in 10 Years© Copyright 2011 EMC Corporation. All rights reserved. 3
  4. 4. Why Disk Aggregation is Losing SteamMoore’s Law drives the escalating need for IO transactions The Number of Drives Needed per Host over Time 262,833 We are 164,682 Here ? 103,184 64,652 40,509 25,381 9,964 15,903 23 36 58 93 148 237 378 603 962 1,536 2,451 3,912 6,243 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018 2019 2020© Copyright 2011 EMC Corporation. All rights reserved. 4
  5. 5. The CPU to HDD Performance GapCPU improves 100 times every decade – disk speed hasn’t CPU continue to improve MOORE’S LAW: while disk drive 100X performance remains flat. As a result, applications per will suffer more and more unless we rapidly move to DECADE FLASH. 10,000 times FLASH improved 100 times improved 2000 2010 2020© Copyright 2011 EMC Corporation. All rights reserved. 5
  6. 6. Anatomy of a Enterprise FLASH DriveDesigned for reliability, data integrity and performance Controller SLC NAND FLASH SAS or SATA ports DRAM End to End CRC© Copyright 2011 EMC Corporation. All rights reserved. 6
  7. 7. Comparing Associated CostsWhich technology is the most efficient? Capacity Acquisition Cost Capacity Power Cost $- $5 $10 $15 $20 $25 $30 0 5 10 15 20 25 30 15K HDD 15K HDD 7200 HDD Lowest Capacity Cost 7200 HDD FLASH FLASH FLASH 7200 HDD 15K HDD FLASH 7200 HDD 15K HDD $/GB $24.75 $0.46 $1.80 mWatt/GB 25 4 28 Transaction Acquisition Cost Transaction Power Cost $- $5 $10 $15 $20 0.0 50.0 100.0 150.0 15K HDD 15K HDD 7200 HDD 7200 HDD FLASH Lowest Transaction Cost FLASH FLASH 7200 HDD 15K HDD FLASH 7200 HDD 15K HDD $/IOPS $0.99 $15.28 $6.00 mWatt/IOPS 1.0 133.3 94.4© Copyright 2011 EMC Corporation. All rights reserved. 7
  8. 8. Two Major Components of Your SLATransaction and Capacity support needed by the business? Lowest $/IO SLA Target SLA SSD Needed Transactions $ Lowest $/GB NL-HDD Needed Capacity© Copyright 2011 EMC Corporation. All rights reserved. 8
  9. 9. FLASH Becomes More AffordableProjected to fall below $1/GB by 2018 40% Consumer electronics keep driving FLASH FAB capacity up. As a result, FLASH LOWER NAND prices will continue to fall EACH YEAR© Copyright 2011 EMC Corporation. All rights reserved. 9
  10. 10. ROI Improvement Decade OutlookFLASH technology will improve faster than mechanical drives 1000.0 100.0 FLASH STORAGE 10.0 29X MECH. STORAGE MORE IMPROVED 1.0 0.1 2010 2011 2012 2013 2014 2015 2016 2017 2018 2019 2020 IOPS per $ 0.1 0.3 0.7 1.6 3.7 8.6 20.2 47.1 109.8 256.3 598.0 GB per $ 0.7 1.1 1.9 3.1 5.1 8.6 14.3 23.8 39.7 66.2 110.3© Copyright 2011 EMC Corporation. All rights reserved. 10
  11. 11. The VNX5500-FHigh performance, high availability FLASH Array© Copyright 2011 EMC Corporation. All rights reserved. 11
  12. 12. The EMC VNX 5500-F Flash ArrayProven high availability and consistent high performance • 5 x 9’s availability for mission critical 10X applications – RAID data protection, proactive global sparing, replication, EMC quality Enterprise Flash TRANSACTIONS • Starter configuration at 2 or 4TB of 1/8 th Flash – 25 X 2.5” drives, with 100GB or 200GB SLC SSDs THE COST • Advanced data efficiency services – Compression, De-dupe, and Thin Provisioning doubles usable capacity • Full unified protocol support – CIFS, NFS, pNFS, iSCSI, FCP and FCoE • Expandable to Tiered Storage with FAST – From 49 TB of Flash to 675 TB of tiered storage (Flash, SAS, NL-SAS) as data ages© Copyright 2011 EMC Corporation. All rights reserved. 12
  13. 13. The FLASH 1st Data Management StrategyHot data on fast FLASH SSDs—cold data on dense disks “Hot” high activity Highly active data is As data ages, activity falls, stored on triggering automatic FLASH SSDs for movement to high fastest capacity disk drives for response time lowest cost Data Activity High FLASH Cap. SSD HDD Movement Trigger “Cold” low activity Data Age (5 years)© Copyright 2011 EMC Corporation. All rights reserved. 13
  14. 14. Gaining Economic EfficiencyFLASH SSDs for active transactions & 3TB HDD for overflow 1.2 • FLASH SSDs offers highest transaction efficiency 1 FLASH SSD • Large capacity HDD offers 0.8 highest capacity efficiency • High RPM HDDs are not capacity- IOPS per $ 0.6 FAST efficient nor are they 0.4 VP performance-efficient • FAST VP automatically and 0.2 15K seamlessly optimizes data 0 HDD 3TB HDD across all disk technologies 0 0.5 1 1.5 2 2.5 GB per $© Copyright 2011 EMC Corporation. All rights reserved. 14
  15. 15. Today’s Compute HierarchyRight data. Right Place. Right Cost. Multi-Core/Socket CPUs pS − pS latency DDR4 - 4.266GHz RAM nS − 7 to 200nS latency FAST Cache 200GB FLASH SSDs uS − 20 to 320 uS latency FAST Virtual Pools 3TB HDDs mS − 7 to 34 mS latency© Copyright 2011 EMC Corporation. All rights reserved. 15
  16. 16. Optimizing Your VNX DeploymentLowest $/IOPS and lowest $/GB Max System IOPS Transactions Add SSDs Add HDDs for more IOPS for more Capacity System Scale© Copyright 2011 EMC Corporation. All rights reserved. 16
  17. 17. Hybrid SystemsA little FLASH goes a long way 90 SSDs HDDs 85 910 70 415 50 180 45 75 30 VNX5100 VNX5300 VNX5500 VNX5700 VNX7500© Copyright 2011 EMC Corporation. All rights reserved. 17
  18. 18. The FAST SuiteDynamically optimizes IO for FLASH 1st at a 64K page size • FAST Cache FAST Cache - FAST VP Relationship – Caches data from the HDD or NL-HDD tiers in Controller the pool – Operates at a page granularity of 64K • FAST VP FAST Cache – Dynamically moves data between tiers in the 64K 64K storage pool – Operates at a slice granularity of 1GB 64K 1G SSD HDD NL-HDD • Deploying both together FAST Virtual Pool ensures maximum IO granularity© Copyright 2011 EMC Corporation. All rights reserved. 18
  19. 19. IO SkewRandom slice activity 1000 1000 900 900 800 800 700 700 600 600 500 500 All 400 400 80% 300 300 200 200 100 100 0 0 1 6 11 16 21 26 31 36 41 46 51 56 61 66 71 76 81 86 91 96 101 1 7 13 19 25 31 37 43 49 55 61 67 73 79 85 91 97• Slice activity 100% randomized • Slice activity stack ranked according to access frequency of a given slice • 51 slices of 101 has 80% of IO© Copyright 2011 EMC Corporation. All rights reserved. 19
  20. 20. IO SkewRandom slice activity weighted by 3% cooling rate 1000 900 800 700 600 • Stack ranked slice map 500 All weighted by a 3% cooling rate 400 "80%" 300 • Now only 31 slices of 101 total 200 has 80% of IO 100 0 1 6 11 16 21 26 31 36 41 46 51 56 61 66 71 76 81 86 91 96 101© Copyright 2011 EMC Corporation. All rights reserved. 20
  21. 21. Examples of IO SkewDriven by Data Growth and Business Models Very Growth: 60% Std. Growth: 50% High Days Hot: 10 Days Hot: 60 Skew 1% of data = 80% of IO Skew 5.5% of data = 80% of IO High Growth: 100% Days Hot: 30 Low Growth: 50% Days Hot: 90 Skew 4.2% of data = 80% of IO Skew 8.2% of data = 80% of IO© Copyright 2011 EMC Corporation. All rights reserved. 21
  22. 22. How Much FLASH?Used dynamically with FAST Size of FLASH FIFO • Data follows a predictable decay in activity Amount of Data being created daily • Older data is constantly being replaced by new highly active data • The amount of FLASH FLASH Capacity required is determined by: – The amount of data created each day, and Number of Days of – The period of time it takes High Data Activity to cool© Copyright 2011 EMC Corporation. All rights reserved. 22
  23. 23. Modeling “Data Decay”A simple GB-Day state model describes typical data behavior P = (1-C) A, and Probability of data access A = LOG1-C (P) Where: P is Probability of data access C is daily Cooling rate A is data Age in days Hot 1GB data age in days© Copyright 2011 EMC Corporation. All rights reserved. 23
  24. 24. Modeling “Data Decay”A simple GB-Day state model describes typical data behavior P = (1-C) A, and A = LOG1-C (P) Hot Where: P is Probability of data access State C is daily Cooling rate Change A is data Age in days Cold Hot Cold 1GB data age in days© Copyright 2011 EMC Corporation. All rights reserved. 24
  25. 25. Basic Data Decay Model Cache FAST • The basic storage pool distribution model assumes FLASH as the first line of defense 20% for 80% of all IOs • As an additional performance buffer, 15K FLASH Tier drives are used for overflow for 19% of all IOs • Only a 1% chance remains that any IOs will be serviced from slow 2TB drives 60% 15K 1% 7 days 52days 250 days© Copyright 2011 EMC Corporation. All rights reserved. 25
  26. 26. How Long Does Data Stay Hot?Depends on business model, applications and workload Daily Hot Days cooling rate @ 80% 30 5.2% 60 2.7% Hot 90 1.8% 120 1.4% Cold 30 days 60 days 90 days 120 days Days of high data activity© Copyright 2011 EMC Corporation. All rights reserved. 26
  27. 27. FLASH PercentageA function of data growth and service level • Relative more FLASH is needed when: – Service level (FLASH hit rate) is elected high, andFLASH portion – Data growth is high • More than 25% FLASH is highly unlikely! Data Growth Rate© Copyright 2011 EMC Corporation. All rights reserved. 27
  28. 28. Getting the Right Blend of Flash3 fundamental business questions • How much data is under management today?  30 TBs • How much is your data growing each year?  50% YoY • How long does your data stay hot?  60 days© Copyright 2011 EMC Corporation. All rights reserved. 28
  29. 29. Calculating Net New DataNet new is a function of amount of starting data and growth 50% Growth Rate • 50% of 30TB is 15TB 15 • The average amount of data generated each day: – 15 X 1024 GB / 365 = 42 GB per day 30 30 Start Year 1© Copyright 2011 EMC Corporation. All rights reserved. 29
  30. 30. Calculating Needed FLASHFLASH only needs to be large enough to hold hot data FLASH portion • FLASH capacity: 2.5 – 60 days X 42 GB = Needed FLASH 2,520 GB 12.5 • FLASH Percentage: – FLASH Capacity/Total Capacity 30 30 – 2,520 GB/(45 X 1024) GB X 100 = 5.4% Start Year 1© Copyright 2011 EMC Corporation. All rights reserved. 30
  31. 31. Calculating How Much FLASH FLASH Yearly Growth Rate% X Number of Hot Days X 100 FLASH % = Portion 365 X (Yearly Growth Rate% + 100%) FLASH portion as a function of Yearly Data Growth 18% 16% of storage pool FLASH portion 14% 12% 10% 8% 6% 4% 2% 0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% 30 days Hot 1% 1% 2% 2% 3% 3% 3% 4% 4% 4% 60 days Hot 1% 3% 4% 5% 5% 6% 7% 7% 8% 8% 90 days Hot 2% 4% 6% 7% 8% 9% 10% 11% 12% 12% 120 days Hot 3% 5% 8% 9% 11% 12% 14% 15% 16% 16%© Copyright 2011 EMC Corporation. All rights reserved. 31
  32. 32. Using the Business Data to ConfigureWith no trace data, the model is used • The SSD tier retains a 80% probability of SSD data re-use • An 15K HDD tier is 3-Tier used as overflow with a 19% chance 60 178 FLASH 1st of data re-use Hot • Finally, the NL-HDD days days Strategy tier is used for long term online storage with a 1% chance and lower of data re- use 15K-HDD Cold NL-HDD© Copyright 2011 EMC Corporation. All rights reserved. 32
  33. 33. Configuration Example30TB usable , 50% growth, 60 days hot @ 80% FLASH service VNX5500 VNX5300 Large Foot Print Smaller Foot Print 200 SAS 15K HDDs SSDs for FAST Cache 4 SSDs for FAST VP 20 15K SAS HDDs for FAST VP 15 NL-SAS HDDs for FAST VP 24 TOTAL 63 $380,573 $182,882 Monolithic© Copyright 2011 EMC Corporation. All rights reserved. 33
  34. 34. Tangible Business BenefitsDramatically lower cost and higher transaction support $ per Usable GB Yearly Power Consumption With FAST $3.97 With FAST 7,030 Without FAST $8.26 Without FAST 30,660 Disk Slots Needed Transactions Supported With FAST 63 With FAST 36,686 Without FAST 200 Without FAST 15,050© Copyright 2011 EMC Corporation. All rights reserved. 34
  35. 35. 12.5X Compounded Improvement in ROI$/GB and $/IO is improves simultaneously ROI Improvement from FLASH 1st 0.25 0.25 GB/$ 0.20 0.20 IO/$ IOPS per Budget $ 0.15 5X 0.10 0.05 0.1 GB/$ 0.04 IO/$ 0.00 0.00 0.05 0.10 0.15 2.5X 0.20 0.25 0.30 GB per Budget $© Copyright 2011 EMC Corporation. All rights reserved. 35
  36. 36. Data Strategy MatrixTechnology differentiated data placement strategy Shared NL HDD Dedicated PCIe FLASH Read Automation Write Shared 15K HDD Shared FLASH SSD Low Activity High Activity© Copyright 2011 EMC Corporation. All rights reserved. 36
  37. 37. Virtualization Further Pressures StorageMoore’s law and better CPU utilization drives IOPS need 32 cores 32 cores 32 cores 32 cores 1M IOPS 1M IOPS 1M IOPS 1M IOPS 2M IOPS • Application management - like vMotion – requires shared storage NEED for 200K • A performance budget of 1M IOPS per host is now a reality IOPS • Shared storage risks becoming SPEED bottleneck in high performance virtualization deployments Note: 2M host pressure IOPS is based on 50% storage side IOPS per host© Copyright 2011 EMC Corporation. All rights reserved. 37
  38. 38. VNX Virtualized Cloud ArchitectureHost-side read caching multiplies performance Performance Scale 2M IOPS 10X • Scale out for “hot data” and read performance with FLASH based host- Capacity Scale side caching • Scale up for deep archive of “cold data” • All writes are protected by BETTER synchronous cache mirroring to DRAM & FLASH in central array PERFORMANCE Note: 2M storage IOPS is based on 90%/10% read/write mix© Copyright 2011 EMC Corporation. All rights reserved. 38
  39. 39. High Performance Storage HierarchyThe data continuum is FLASH centric and extends to the host • Host-side read FLASH caching shields data from Working Set 50uS SAN and array latency PCIe FLASH • Array-side FLASH protects data against host failure SSD 200uS • Array-side HDDs FLASH accumulates inactive data at lowest cost Inactive Data 1,500uS HDD© Copyright 2011 EMC Corporation. All rights reserved. 39
  40. 40. The IO Spread: Access DensityPCIe FLASH offers 8,333X better data access than 15K HDD 10000 1000 100 10 2,500 1 25 0.1 0.3 0.01 0.03 NL-HDD HDD SSD PCIe FLASH Access Density: IOPS/GB© Copyright 2011 EMC Corporation. All rights reserved. 40
  41. 41. Data ContinuumAs data activity falls, different technologies apply “Hot” high activity VFCache Data Activity SSD Dynamic Storage FAST Cache COST Hierarchy differentiated OPTIMIZED SSD on Access Density Cost FAST VP 15K HDD “Cold” NL-HDD low activity Data Age© Copyright 2011 EMC Corporation. All rights reserved. 41
  42. 42. Storage Efficiencies/Flash SessionsUSD Lectures- Session Name Day & Time Monday 10:00am - 11:00am#7 VNX Storage Efficiencies – What, Why, and When Tuesday 11:30am - 12:30pm VNX Virtual Provisioning - Leveraging the Best in Efficiency and Simplicity with Virtual Pool Tuesday 10:00am - 11:00am#8 Provisioning Wednesday 11:30am - 12:30pm Tuesday 4:15pm - 5:15pm#9 VNX Compression and Deduplication - Optimizing Capacity for Your File and Block Infrastructure Thursday 10:00am - 11:00am Tuesday 11:30am - 12:30pm#10 VNX FAST VP - Optimizing Performance and Utilization of Virtual Pools Thursday 8:30am - 9:30am Monday 4:00pm - 5:00pm#11 VNX FAST Cache – Super Charge your Storage with Extended Cache Wednesday 10:00am - 11:00am Tuesday 2:45pm - 3:45pm#12 FAST Storage Design Basics for EMC VNX Wednesday 4:15 - 5:15pm Wednesday 11:30am - 12:30pm#13 Leveraging SSD: Designing for FAST Cache and FAST VP on Unified Block Storage Thursday 11;30am - 12:30pm Monday 8:30am - 9:30am#25 FLASH 1st - The Storage Strategy for the Next Decade Monday 2:30pm - 3:30pm Monday 8:30am - 9:30amFBU#1 Introduction To EMC VFCache Wednesday 4:15pm - 5:15pm Monday 4:00pm – 5:00pmFBU#2 Leveraging EMC VFCache with Enterprise Applications Thursday 10:00am - 11:00amBOF Dive into EMC VNX & FAST Suite Storage Efficiencies Tuesday 1:30pm - 2:30pmBOF Flash Technology in the Storage Environment Wednesday 1:30pm - 2:30pm © Copyright 2011 EMC Corporation. All rights reserved. 42
  43. 43. Provide Feedback & Win! • 125 attendees will receive $100 iTunes gift cards. To enter the raffle, simply complete: – 5 sessions surveys – The conference survey • Download the EMC World Conference App to learn more: emcworld.com/app© Copyright 2011 EMC Corporation. All rights reserved. 43
  44. 44. © Copyright 2011 EMC Corporation. All rights reserved. 44
  45. 45. FLASH 1st THANK YOU Denis Vilfort Sr. Director, USD Marketing© Copyright 2011 EMC Corporation. All rights reserved. 45

×