AN ALTERNATIVE TO ALL- FLASH ARRAYS: PREDICTIVE STORAGE CACHING - PowerPoint PPT Presentation

AN ALTERNATIVE TO ALL- FLASH ARRAYS: PREDICTIVE STORAGE CACHING THE EASIEST WAY TO INCREASE PERFORMANCE AND LOWER STORAGE COSTS Bruce Kornfeld , Chief Marketing Officer, StorMagic Luke Pruen , Technical Services Director, StorMagic Peter Smith , System Admin, Harris Corporation

STORMAGIC SVSAN: BRIEF OVERVIEW SvSAN turns the internal disk, SSD and memory of two or more servers into highly available shared storage. 2

ALL-FLASH ARRAYS ARE TEMPTING • SSDs vs. HDD • 10x-20x more performance • 10x more expensive • Lots of hype from flash array vendors • Not all workloads need all-flash • A balanced approach with Source: Statistica advanced caching algorithms SSDs are still 10x more costly than HDD. can meet your needs SSD: $0.50 - $1.00 per GB HDD: $0.05 - $0.10 per GB 3

PREDICTIVE STORAGE CACHING Over 400% performance improvement with a patent-pending method for predicting when data will become ‘hot’ AUTOMATED BUILT FOR COST EFFECTIVE PERFORMANCE Two caches – system Use fewer and less expensive memory and SSD disk drives (7.2K) Lower latency – less waiting for spinning disks Patent-pending predictive Lower CAPEX – server algorithms memory is inexpensive Data pinning Automation tracks ‘hot’ data Lower OPEX – power, cooling Solves the virtual server ‘I/O and places in the right cache and maintenance blender effect’ 4

AN ALTERNATIVE TO ALL- FLASH ARRAYS: PREDICTIVE STORAGE CACHING THE EASIEST WAY TO INCREASE PERFORMANCE AND LOWER STORAGE COSTS Luke Pruen , Technical Services Director, StorMagic

TODAY’S STORAGE OPTIONS • The performance gap between CPU and storage • Disk only • High capacity, low performance workloads • All-flash • Performance at any cost workloads • Hybrid • Most workloads fit here 6

THE IMPORTANCE OF CACHING • Virtualized environments suffer from the ‘I/O blender effect’ • Working sets of data change over time • Advanced caching can solve both problems 7

WRITE CACHING Stage 1 • All new data written to SSD • Data marked ‘dirty’ – not committed to HDDs Stage 2 • Write operation is acknowledged immediately to the server Stage 3 • ‘Dirty’ data is reordered and grouped based on disk locality • Data de-staged and written to HDD, sequentially as possible Stage 4 • SSD cache notified when data successfully written to HDD • Cached data marked ‘clean’, remains in cache until space needed 8

WRITE CACHING Stage 1 • All new data written to SSD • Data marked ‘dirty’ – not committed to HDDs Stage 2 • Write operation is acknowledged immediately to the server Stage 3 • ‘Dirty’ data is reordered and grouped based on disk locality • Data de-staged and written to HDD, sequentially as possible Stage 4 • SSD cache notified when data successfully written to HDD • Cached data marked ‘clean’, remains in cache until space needed 8

READ AHEAD & DATA PINNING Read ahead mode Data pinning mode • I/O blender effect aware! • Pin specific data/workloads in memory • Identifies sequential interleaved I/Os • Delivers most efficient read performance • Detects sequential read streams • DBs, VDI, frequently repeated operations • Pre-fetches data into memory • Manage multiple pin groups 9

PREDICTIVE CACHING Advanced, patent-pending method for predicting ‘hot’ data, placing it on the best storage cache available Predictive approach Sizing • Tracker module has 7 levels • Assign cache sizes to meet requirements • All read I/Os monitored and analyzed • Grow caches as working sets change • Most frequently used – higher levels • Cache placement based on levels • Use any combination of memory, SSD/flash and disk Flexible storage options Play to the strengths • RAM: Most frequently accessed data • Play to the strengths of all mediums • SSD/flash: Next most frequently • Memory highest IOPS accessed data • SSD/flash magnetic drives providing • HDD: Infrequently accessed data – low price per GB ‘cold’ data 10

REAL CUSTOMER DATA (BEFORE CACHING) Performance characteristics of their existing workloads – measured by StorMagic Block Size Distribution Read Write 25,000,000 Workload 20,000,000 77% 23% Read/Write % Hit Count 15,000,000 Sequential % 49% 39% Writes • 12 Virtual Machines 10,000,000 Average Per Day 991 GB 294 GB Read • 78 applications 5,000,000 Average Block Size 58 KB 54 KB 0 212 138 Average IOPS 1 2 4 8 16 32 64 128 256 512 1 2 4 KB KB KB KB KB KB KB KB KB KB MB MB MB Block Size Throughput IOPs Locality of access 3500 10000 Number of accesses 3000 (logarithmic scale) 1000 2500 Thousands 100 2000 10 IOPs 1 1500 Read Read 0.1 Write Write 1000 0.01 500 0.001 0 21 MB 110 GB 221 GB 332 GB 443 GB 554 GB 665 GB 776 GB 887 GB 998 GB 1.1 TB 1.2 TB 1.3 TB 1.4 TB 1.5 TB 1.6 TB 1.7 TB 1.8 TB 2.0 TB 2.1 TB 2.2 TB 2.3 TB 2.4 TB 2.5 TB 2.6 TB 18:39 21:25 00:11 02:57 05:43 08:29 11:15 14:01 16:47 19:33 22:19 01:05 03:51 06:37 09:23 12:09 14:55 17:41 20:27 23:13 01:59 04:45 07:31 10:17 13:03 15:49 18:35 Time of Day (UTC) 11

REAL CUSTOMER DATA (WITH CACHING) Impact of Predictive Storage Caching to increase performance HDD Only (no caching) HDD & SSD (with caching) Total IOPS 0 2000 4000 6000 8000 10000 12000 14000 • 1 x RAID5 = 3 x 1.2TB 10K SAS • 1 x 200GB Samsung SSD disks • 1 x RAID5 = 3 x 1.2TB 10K SAS HDD Only 2400 disks HDD & Memory 12341 HDD & Memory (with caching) HDD, SSD & Memory (with caching) HDD & SSD 6165 • 12GB of memory per host for • 12GB of memory per host for caching caching HDD, SSD & Memory 13061 • 1 x RAID5 = 3 x 1.2TB 10K SAS • 1 x 200GB Samsung SSD disks • 1 x RAID5 = 3 x 1.2TB 10K SAS AVG Latency (ms) disks 0 5 10 15 20 25 HDD Only 21.86 With Caching HDD & Memory 4.94 HDD Only HDD & Memory HDD & SSD HDD, SSD & Memory Total IOPS 2400 12341 6165 13061 HDD & SSD 9.37 21.86 4.94 9.37 5.21 AVG Latency (ms) HDD, SSD & Memory 5.21 12

SYNTHETIC ‘HERO’ NUMBERS High performance achievement in a controlled lab environment IOPS Test HDD Config 4263 4KB 100% RND RD 207117 • vSphere 6.5 • 1 x RAID5 = 3 x 1.2TB 10K SAS disks 2004 4KB 100% RND WR • 2 x Windows VMs Tiered Config 38196 HDD 1832 • IOmeter • 1 x RAID5 = 3 x 1.2TB 10K SAS disks 4KB 70/30 RD/WR 80% RND 79738 Tiered • VMDK on VMFS • 1 x 200GB Samsung SSD 7729 32KB 100% SEQ RD 85803 • 32GB of memory per host for caching 8728 32KB 100% SEQ WR 15651 AVG Latency (ms) Total IOPS Average Response (ms) 60.04 4KB 100% RND RD Workload HDD Tiered HDD Tiered 1.24 4KB 100% Random Read 4263 207117 60.04 1.24 127.17 4KB 100% RND WR 2004 38196 127.17 6.70 4KB 100% Random Write 6.70 1832 79738 147.36 3.21 4KB 70/30 Read Write 80% Random 147.36 HDD 4KB 70/30 RD/WR 80% RND 32KB 100% Sequential Read 7729 85803 33.62 2.98 3.21 Tiered 32KB 100% Sequential Write 8728 15651 29.33 16.36 33.62 32KB 100% SEQ RD 2.98 29.33 32KB 100% SEQ WR 13 16.36

SUMMARY: PREDICTIVE STORAGE CACHING Increased performance – without the high cost of all-flash arrays Flexible configurations – we’ll help ‘right-size’ the server hardware Lower CAPEX – use slower drives for capacity, low-cost system memory and some SSD for performance Lower OPEX – less power, cooling and maintenance 14

Q&A – HARRIS CORPORATION Peter Smith Systems Administrator Network, Platforms and Security Group Harris Corporation 15

NEXT STEPS Further reading: Download your free An overview of SvSAN – http://stormagic.com/svsan/ trial of SvSAN SvSAN data sheet – http://stormagic.com/svsan-data-sheet/ stormagic.com/trial Predictive Storage Caching – http://stormagic.com/svsan/predictive-storage-caching/ SvSAN Product Information Product Options SvSAN license 2, 6, 12 and unlimited TBs License entitlement 2 mirrored servers Maintenance and support Platinum - 24x7 / Gold - 9x5 For further information, please contact: sales@stormagic.com 16