SlideShare una empresa de Scribd logo
1 de 31
1© Copyright 2009 EMC Corporation. All rights reserved.
Fast Track Foundations:
Serious about Sequential I/O
James Rowland-Jones
&
Allan Mitchell
2© Copyright 2009 EMC Corporation. All rights reserved.
WTF IS FAST TRACK?
Introduction
3© Copyright 2009 EMC Corporation. All rights reserved.
Blend of Hardware & Software
4© Copyright 2009 EMC Corporation. All rights reserved.
Predictable Performance
Balanced
Configurations
– CPU
– Memory
– Network
– Storage
5© Copyright 2009 EMC Corporation. All rights reserved.
Targeted to the Data Warehouse Workload
Sequential Scanning of Data
Leads to fewer disks
6© Copyright 2009 EMC Corporation. All rights reserved.
Prescriptive
Best Practices
Recommendations
You don’t get to
pick and choose!
7© Copyright 2009 EMC Corporation. All rights reserved.
WHY SQL 2008 ONLY?
8© Copyright 2009 EMC Corporation. All rights reserved.
SQL Server 2008 Enterprise...
...was the Data Warehousing release...
Compression
Star Joins
Few Outer Rows Parallelism
Partitioned Table Parallelism
Resource Governor
Improved Read Ahead
9© Copyright 2009 EMC Corporation. All rights reserved.
SQL 2008 Settings
Startup Parameters
– T1117
– E
Resource Governor
– Reduce Max memory per session
Page Compression ON!
10© Copyright 2009 EMC Corporation. All rights reserved.
Windows 2008 Settings
Multipath I/O Feature is required
Lock Pages In Memory Enabled
11© Copyright 2009 EMC Corporation. All rights reserved.
SEQUENTIAL VS RANDOM I/O
DEMO TIME:
12© Copyright 2009 EMC Corporation. All rights reserved.
Counters
Performance Counter
 Avg. Disk Bytes / Read
 Avg. Disk Sec / Read
 Disk Read Bytes / Sec
 Disk Reads / Sec
Meaning
 Size of I/Os
 Latency “Seek Time”
 Throughput – MB/s
 I/Os per second
 sys.dm_db_index_physical_file_stats
 sys.dm_io_virtual_file_stats
 sys.dm_io_pending_io_requests
DMVs
13© Copyright 2009 EMC Corporation. All rights reserved.
Time to Take the Red Pill...
14© Copyright 2009 EMC Corporation. All rights reserved.
15© Copyright 2009 EMC Corporation. All rights reserved.
DB DESIGN CONSIDERATIONS
Fast Track Thinking
16© Copyright 2009 EMC Corporation. All rights reserved.
Indexing
Think Index Light
 NC Indexes aimed at low volume row lookups
 Indexes are Expensive to maintain
17© Copyright 2009 EMC Corporation. All rights reserved.
Aggregation
Think Aggregation Light
 Introduces Updates (fragmentation)
 Removing aggs reduces maintenance
18© Copyright 2009 EMC Corporation. All rights reserved.
Partitioning
Partition Qualification
Data Management
P1 P4P3P2P
2
19© Copyright 2009 EMC Corporation. All rights reserved.
RG1 RG2
RG3 RG4
RG5 RG6
RG7 RG8
RG9
RG10
Conceptual View of Fast Track Hardware
1
3
2
4
5
7
6
8
H
H
4GB
4GB
4GB
4GB
4GB
4GB
4GB
RAID 1
LUN 01
LUN 02
LUN 03
LUN 04
LUN 17
(Logs)
LUN 05
LUN 07
LUN 06
LUN 08
SPARE
SPASPB
FC1
FC2
FC3
FC4
FibreSwitch
LUN 09
LUN 11
LUN 10
LUN 12
LUN 18
(Logs)
LUN 13
LUN 15
LUN 14
LUN 16
SPARE
SPCSPD
FC5
FC6
FC7
FC8
8 Core 32 GB RAM
Server
4GB
20© Copyright 2009 EMC Corporation. All rights reserved.
RG1 RG2
RG3 RG4
RG5 RG6
RG7 RG8
RG9
RG10
Performance View of Fast Track Hardware
1
3
2
4
5
7
6
8
H
H
LUN 01
LUN 02
LUN 03
LUN 04
LUN 17
(Logs)
LUN 05
LUN 07
LUN 06
LUN 08
SPASPB
FC1
FC2
FC3
FC4
FibreSwitch
LUN 09
LUN 11
LUN 10
LUN 12
LUN 18
(Logs)
LUN 13
LUN 15
LUN 14
LUN 16
SPCSPD
FC5
FC6
FC7
FC8
200 mb/s per CPU core is
central to Fast Track
Architecture
32GB
RAM800mb/s
800mb/s
400mb/s
per port
1 LUN =
125
MB/S
SP Port
650MB/s
FCPort
400MB/S
21© Copyright 2009 EMC Corporation. All rights reserved.
Performance Summary
Component Throughput
8 CPU Cores 1600 MB/S
4 HBA Ports (2 adapters) 1600 MB/S
8 FC Ports (4 per array) 3200 MB/S
4 Storage Processors (2 per array) 2600 MB/S
16 LUNS (Data Only) 2000 MB/S
22© Copyright 2009 EMC Corporation. All rights reserved.
Capacity Summary
16x 300GB Data Disks Raw (formatted) 4352 GB
4x 300GB Log Disks Raw (formatted) 1088 GB
LESS RAID 1
8x RAID 1 Raid Groups for Data 2176 GB
2x RAID 1 Raid Groups for Logs 544 GB
LESS TEMPDB SPACE
Tempdb Data files 576 GB
User Data Uncompressed Available 1600 GB
Data Compression Ratio 2.5:1 4000 GB
23© Copyright 2009 EMC Corporation. All rights reserved.
Database Layout
Fi1 Fi2 Fi3 Fi4 Fi5 Fi6 Fi7
Fi9 Fi10 Fi11 Fi12 Fi13 Fi14 Fi15
Fi1 Fi2 Fi3 Fi4 Fi5 Fi6 Fi7
Fi1 Fi2 Fi3 Fi4 Fi5 Fi6 Fi7
Fi9 Fi10 Fi11 Fi12 Fi13 Fi14 Fi15
Fi9 Fi10 Fi11 Fi12 Fi13 Fi14 Fi15 tempLog
Live Log
Stage Log
Live
FG1
Stage
FG1
Tempdb
Primary
Fi8
Fi8
Fi8
Fi16
Fi16
Fi16
Fi0
Fi0
PrimaryPrimary
24© Copyright 2009 EMC Corporation. All rights reserved.
MANAGING FRAGMENTATION
The arch enemy of Sequential I/O
25© Copyright 2009 EMC Corporation. All rights reserved.
Fragmentation
File System Extent Index
26© Copyright 2009 EMC Corporation. All rights reserved.
File System Fragmentation
Pre-Allocate your Database files
Avoid O/S de-fraggers
They know not what they do...
27© Copyright 2009 EMC Corporation. All rights reserved.
Extent Fragmentation
Occurs Courtesy of
Concurrent DML
Updates
Deletes
Follow the guidelines!
Otherwise you’ll be re-loading the tables....
28© Copyright 2009 EMC Corporation. All rights reserved.
Index Fragmentation
 Don’t use ALTER INDEX REORGANIZE
 Do one of the following
– ALTER INDEX REBUILD ...
SORT_IN_TEMPDB = TRUE (MAXDOP 1)
– INSERT ... SELECT into a new table
29© Copyright 2009 EMC Corporation. All rights reserved.
File Groups
Isolate Volatile Data
 Create a separate Filegroup
 One file per LUN to spread the I/O
Isolate Small / Dimension Tables
Stage Data in a Separate DB or File Group
30© Copyright 2009 EMC Corporation. All rights reserved.
DATA LOADING FOR
SEQUENTIAL I/O
Over to Mitchell...
where Allan and Darren live

Más contenido relacionado

La actualidad más candente

Intro to Cell Broadband Engine for HPC
Intro to Cell Broadband Engine for HPCIntro to Cell Broadband Engine for HPC
Intro to Cell Broadband Engine for HPCSlide_N
 
Solaris 11 Consolidation Tools
Solaris 11 Consolidation ToolsSolaris 11 Consolidation Tools
Solaris 11 Consolidation ToolsRoman Ivanov
 
Advanced Site Recovery -- Technical Briefing
Advanced Site Recovery -- Technical BriefingAdvanced Site Recovery -- Technical Briefing
Advanced Site Recovery -- Technical BriefingJames Price
 
Resume_CQ_Edward
Resume_CQ_EdwardResume_CQ_Edward
Resume_CQ_Edwardcaiqi wang
 
OSインストーラーの自作方法
OSインストーラーの自作方法OSインストーラーの自作方法
OSインストーラーの自作方法LINE Corporation
 
Presentation vmax hardware deep dive
Presentation   vmax hardware deep divePresentation   vmax hardware deep dive
Presentation vmax hardware deep divesolarisyougood
 
Storage Area Networking: SAN Technology Update & Best Practice Deep Dive for ...
Storage Area Networking: SAN Technology Update & Best Practice Deep Dive for ...Storage Area Networking: SAN Technology Update & Best Practice Deep Dive for ...
Storage Area Networking: SAN Technology Update & Best Practice Deep Dive for ...EMC
 
M. Rafaat_EMC_Presentation
M. Rafaat_EMC_PresentationM. Rafaat_EMC_Presentation
M. Rafaat_EMC_PresentationMohamed Raafat
 
A z/OS System Programmer’s Guide to Migrating to a New IBM System z9 EC or z9...
A z/OS System Programmer’s Guide to Migrating to a New IBM System z9 EC or z9...A z/OS System Programmer’s Guide to Migrating to a New IBM System z9 EC or z9...
A z/OS System Programmer’s Guide to Migrating to a New IBM System z9 EC or z9...IBM India Smarter Computing
 
ASCI Terascale Simulation Requirements and Deployments
ASCI Terascale Simulation Requirements and DeploymentsASCI Terascale Simulation Requirements and Deployments
ASCI Terascale Simulation Requirements and DeploymentsGlenn K. Lockwood
 
Mega Launch Recap Slide Deck
Mega Launch Recap Slide DeckMega Launch Recap Slide Deck
Mega Launch Recap Slide DeckVarrow Inc.
 
Presentation symmetrix vmax family with enginuity 5876
Presentation   symmetrix vmax family with enginuity 5876Presentation   symmetrix vmax family with enginuity 5876
Presentation symmetrix vmax family with enginuity 5876solarisyougood
 

La actualidad más candente (13)

Intro to Cell Broadband Engine for HPC
Intro to Cell Broadband Engine for HPCIntro to Cell Broadband Engine for HPC
Intro to Cell Broadband Engine for HPC
 
Solaris 11 Consolidation Tools
Solaris 11 Consolidation ToolsSolaris 11 Consolidation Tools
Solaris 11 Consolidation Tools
 
Advanced Site Recovery -- Technical Briefing
Advanced Site Recovery -- Technical BriefingAdvanced Site Recovery -- Technical Briefing
Advanced Site Recovery -- Technical Briefing
 
Resume_CQ_Edward
Resume_CQ_EdwardResume_CQ_Edward
Resume_CQ_Edward
 
OSインストーラーの自作方法
OSインストーラーの自作方法OSインストーラーの自作方法
OSインストーラーの自作方法
 
Presentation vmax hardware deep dive
Presentation   vmax hardware deep divePresentation   vmax hardware deep dive
Presentation vmax hardware deep dive
 
Storage Area Networking: SAN Technology Update & Best Practice Deep Dive for ...
Storage Area Networking: SAN Technology Update & Best Practice Deep Dive for ...Storage Area Networking: SAN Technology Update & Best Practice Deep Dive for ...
Storage Area Networking: SAN Technology Update & Best Practice Deep Dive for ...
 
M. Rafaat_EMC_Presentation
M. Rafaat_EMC_PresentationM. Rafaat_EMC_Presentation
M. Rafaat_EMC_Presentation
 
A z/OS System Programmer’s Guide to Migrating to a New IBM System z9 EC or z9...
A z/OS System Programmer’s Guide to Migrating to a New IBM System z9 EC or z9...A z/OS System Programmer’s Guide to Migrating to a New IBM System z9 EC or z9...
A z/OS System Programmer’s Guide to Migrating to a New IBM System z9 EC or z9...
 
ASCI Terascale Simulation Requirements and Deployments
ASCI Terascale Simulation Requirements and DeploymentsASCI Terascale Simulation Requirements and Deployments
ASCI Terascale Simulation Requirements and Deployments
 
Mega Launch Recap Slide Deck
Mega Launch Recap Slide DeckMega Launch Recap Slide Deck
Mega Launch Recap Slide Deck
 
EMCSymmetrix vmax-10
EMCSymmetrix vmax-10EMCSymmetrix vmax-10
EMCSymmetrix vmax-10
 
Presentation symmetrix vmax family with enginuity 5876
Presentation   symmetrix vmax family with enginuity 5876Presentation   symmetrix vmax family with enginuity 5876
Presentation symmetrix vmax family with enginuity 5876
 

Similar a Fast track foundations getting serious about sequential io

Transforming your Business with Scale-Out Flash: How MongoDB & Flash Accelera...
Transforming your Business with Scale-Out Flash: How MongoDB & Flash Accelera...Transforming your Business with Scale-Out Flash: How MongoDB & Flash Accelera...
Transforming your Business with Scale-Out Flash: How MongoDB & Flash Accelera...MongoDB
 
Z109889 z4 r-storage-dfsms-vegas-v1910b
Z109889 z4 r-storage-dfsms-vegas-v1910bZ109889 z4 r-storage-dfsms-vegas-v1910b
Z109889 z4 r-storage-dfsms-vegas-v1910bTony Pearson
 
Enterprise power systems transition to power7 technology
Enterprise power systems transition to power7 technologyEnterprise power systems transition to power7 technology
Enterprise power systems transition to power7 technologysolarisyougood
 
Presentation power vm common 2012
Presentation   power vm common 2012Presentation   power vm common 2012
Presentation power vm common 2012solarisyougood
 
Z4R: Intro to Storage and DFSMS for z/OS
Z4R: Intro to Storage and DFSMS for z/OSZ4R: Intro to Storage and DFSMS for z/OS
Z4R: Intro to Storage and DFSMS for z/OSTony Pearson
 
Z109889 z4 r-storage-dfsms-jburg-v1909d
Z109889 z4 r-storage-dfsms-jburg-v1909dZ109889 z4 r-storage-dfsms-jburg-v1909d
Z109889 z4 r-storage-dfsms-jburg-v1909dTony Pearson
 
G108277 ds8000-resiliency-lagos-v1905c
G108277 ds8000-resiliency-lagos-v1905cG108277 ds8000-resiliency-lagos-v1905c
G108277 ds8000-resiliency-lagos-v1905cTony Pearson
 
OSBConf 2015 | Contemporary and cost efficient backups to to tape by josef we...
OSBConf 2015 | Contemporary and cost efficient backups to to tape by josef we...OSBConf 2015 | Contemporary and cost efficient backups to to tape by josef we...
OSBConf 2015 | Contemporary and cost efficient backups to to tape by josef we...NETWAYS
 
#IBMEdge: Flash Storage Session
#IBMEdge: Flash Storage Session#IBMEdge: Flash Storage Session
#IBMEdge: Flash Storage SessionBrocade
 
All Flash is not Equal: Tony Pearson contrasts IBM FlashSystem with Solid-Sta...
All Flash is not Equal: Tony Pearson contrasts IBM FlashSystem with Solid-Sta...All Flash is not Equal: Tony Pearson contrasts IBM FlashSystem with Solid-Sta...
All Flash is not Equal: Tony Pearson contrasts IBM FlashSystem with Solid-Sta...Tony Pearson
 
Sparc m6 32 in-memory infrastructure for the entire enterprise
Sparc m6 32 in-memory infrastructure for the entire enterpriseSparc m6 32 in-memory infrastructure for the entire enterprise
Sparc m6 32 in-memory infrastructure for the entire enterprisesolarisyougood
 
Firebird2.5 Benchmarks(English)20091031
Firebird2.5 Benchmarks(English)20091031Firebird2.5 Benchmarks(English)20091031
Firebird2.5 Benchmarks(English)20091031Tsutomu Hayashi
 
DEFCON 23 - Etienne Martineau - inter vm data exfiltration
DEFCON 23 - Etienne Martineau - inter vm data exfiltrationDEFCON 23 - Etienne Martineau - inter vm data exfiltration
DEFCON 23 - Etienne Martineau - inter vm data exfiltrationFelipe Prado
 
Approaching hyperconvergedopenstack
Approaching hyperconvergedopenstackApproaching hyperconvergedopenstack
Approaching hyperconvergedopenstackIkuo Kumagai
 
Verification of the QorIQ Communication Platform Containing CoreNet Fabric wi...
Verification of the QorIQ Communication Platform Containing CoreNet Fabric wi...Verification of the QorIQ Communication Platform Containing CoreNet Fabric wi...
Verification of the QorIQ Communication Platform Containing CoreNet Fabric wi...DVClub
 
Power 7 Overview
Power 7 OverviewPower 7 Overview
Power 7 Overviewlambertt
 
3. EMC Storage for future Surveillance.pdf
3. EMC Storage for future Surveillance.pdf3. EMC Storage for future Surveillance.pdf
3. EMC Storage for future Surveillance.pdfPawachMetharattanara
 
[db tech showcase Tokyo 2018] #dbts2018 #B17 『オラクル パフォーマンス チューニング - 神話、伝説と解決策』
[db tech showcase Tokyo 2018] #dbts2018 #B17 『オラクル パフォーマンス チューニング - 神話、伝説と解決策』[db tech showcase Tokyo 2018] #dbts2018 #B17 『オラクル パフォーマンス チューニング - 神話、伝説と解決策』
[db tech showcase Tokyo 2018] #dbts2018 #B17 『オラクル パフォーマンス チューニング - 神話、伝説と解決策』Insight Technology, Inc.
 
Ceph Day KL - Delivering cost-effective, high performance Ceph cluster
Ceph Day KL - Delivering cost-effective, high performance Ceph clusterCeph Day KL - Delivering cost-effective, high performance Ceph cluster
Ceph Day KL - Delivering cost-effective, high performance Ceph clusterCeph Community
 

Similar a Fast track foundations getting serious about sequential io (20)

Transforming your Business with Scale-Out Flash: How MongoDB & Flash Accelera...
Transforming your Business with Scale-Out Flash: How MongoDB & Flash Accelera...Transforming your Business with Scale-Out Flash: How MongoDB & Flash Accelera...
Transforming your Business with Scale-Out Flash: How MongoDB & Flash Accelera...
 
Z109889 z4 r-storage-dfsms-vegas-v1910b
Z109889 z4 r-storage-dfsms-vegas-v1910bZ109889 z4 r-storage-dfsms-vegas-v1910b
Z109889 z4 r-storage-dfsms-vegas-v1910b
 
Enterprise power systems transition to power7 technology
Enterprise power systems transition to power7 technologyEnterprise power systems transition to power7 technology
Enterprise power systems transition to power7 technology
 
Presentation power vm common 2012
Presentation   power vm common 2012Presentation   power vm common 2012
Presentation power vm common 2012
 
Z4R: Intro to Storage and DFSMS for z/OS
Z4R: Intro to Storage and DFSMS for z/OSZ4R: Intro to Storage and DFSMS for z/OS
Z4R: Intro to Storage and DFSMS for z/OS
 
Z109889 z4 r-storage-dfsms-jburg-v1909d
Z109889 z4 r-storage-dfsms-jburg-v1909dZ109889 z4 r-storage-dfsms-jburg-v1909d
Z109889 z4 r-storage-dfsms-jburg-v1909d
 
G108277 ds8000-resiliency-lagos-v1905c
G108277 ds8000-resiliency-lagos-v1905cG108277 ds8000-resiliency-lagos-v1905c
G108277 ds8000-resiliency-lagos-v1905c
 
OSBConf 2015 | Contemporary and cost efficient backups to to tape by josef we...
OSBConf 2015 | Contemporary and cost efficient backups to to tape by josef we...OSBConf 2015 | Contemporary and cost efficient backups to to tape by josef we...
OSBConf 2015 | Contemporary and cost efficient backups to to tape by josef we...
 
#IBMEdge: Flash Storage Session
#IBMEdge: Flash Storage Session#IBMEdge: Flash Storage Session
#IBMEdge: Flash Storage Session
 
All Flash is not Equal: Tony Pearson contrasts IBM FlashSystem with Solid-Sta...
All Flash is not Equal: Tony Pearson contrasts IBM FlashSystem with Solid-Sta...All Flash is not Equal: Tony Pearson contrasts IBM FlashSystem with Solid-Sta...
All Flash is not Equal: Tony Pearson contrasts IBM FlashSystem with Solid-Sta...
 
Sparc m6 32 in-memory infrastructure for the entire enterprise
Sparc m6 32 in-memory infrastructure for the entire enterpriseSparc m6 32 in-memory infrastructure for the entire enterprise
Sparc m6 32 in-memory infrastructure for the entire enterprise
 
Firebird2.5 Benchmarks(English)20091031
Firebird2.5 Benchmarks(English)20091031Firebird2.5 Benchmarks(English)20091031
Firebird2.5 Benchmarks(English)20091031
 
DEFCON 23 - Etienne Martineau - inter vm data exfiltration
DEFCON 23 - Etienne Martineau - inter vm data exfiltrationDEFCON 23 - Etienne Martineau - inter vm data exfiltration
DEFCON 23 - Etienne Martineau - inter vm data exfiltration
 
Approaching hyperconvergedopenstack
Approaching hyperconvergedopenstackApproaching hyperconvergedopenstack
Approaching hyperconvergedopenstack
 
Verification of the QorIQ Communication Platform Containing CoreNet Fabric wi...
Verification of the QorIQ Communication Platform Containing CoreNet Fabric wi...Verification of the QorIQ Communication Platform Containing CoreNet Fabric wi...
Verification of the QorIQ Communication Platform Containing CoreNet Fabric wi...
 
Power 7 Overview
Power 7 OverviewPower 7 Overview
Power 7 Overview
 
3. EMC Storage for future Surveillance.pdf
3. EMC Storage for future Surveillance.pdf3. EMC Storage for future Surveillance.pdf
3. EMC Storage for future Surveillance.pdf
 
[db tech showcase Tokyo 2018] #dbts2018 #B17 『オラクル パフォーマンス チューニング - 神話、伝説と解決策』
[db tech showcase Tokyo 2018] #dbts2018 #B17 『オラクル パフォーマンス チューニング - 神話、伝説と解決策』[db tech showcase Tokyo 2018] #dbts2018 #B17 『オラクル パフォーマンス チューニング - 神話、伝説と解決策』
[db tech showcase Tokyo 2018] #dbts2018 #B17 『オラクル パフォーマンス チューニング - 神話、伝説と解決策』
 
Ceph Day KL - Delivering cost-effective, high performance Ceph cluster
Ceph Day KL - Delivering cost-effective, high performance Ceph clusterCeph Day KL - Delivering cost-effective, high performance Ceph cluster
Ceph Day KL - Delivering cost-effective, high performance Ceph cluster
 
Ibm cell
Ibm cell Ibm cell
Ibm cell
 

Último

Easier, Faster, and More Powerful – Alles Neu macht der Mai -Wir durchleuchte...
Easier, Faster, and More Powerful – Alles Neu macht der Mai -Wir durchleuchte...Easier, Faster, and More Powerful – Alles Neu macht der Mai -Wir durchleuchte...
Easier, Faster, and More Powerful – Alles Neu macht der Mai -Wir durchleuchte...panagenda
 
IESVE for Early Stage Design and Planning
IESVE for Early Stage Design and PlanningIESVE for Early Stage Design and Planning
IESVE for Early Stage Design and PlanningIES VE
 
ECS 2024 Teams Premium - Pretty Secure
ECS 2024   Teams Premium - Pretty SecureECS 2024   Teams Premium - Pretty Secure
ECS 2024 Teams Premium - Pretty SecureFemke de Vroome
 
Linux Foundation Edge _ Overview of FDO Software Components _ Randy at Intel.pdf
Linux Foundation Edge _ Overview of FDO Software Components _ Randy at Intel.pdfLinux Foundation Edge _ Overview of FDO Software Components _ Randy at Intel.pdf
Linux Foundation Edge _ Overview of FDO Software Components _ Randy at Intel.pdfFIDO Alliance
 
The Metaverse: Are We There Yet?
The  Metaverse:    Are   We  There  Yet?The  Metaverse:    Are   We  There  Yet?
The Metaverse: Are We There Yet?Mark Billinghurst
 
Salesforce Adoption – Metrics, Methods, and Motivation, Antone Kom
Salesforce Adoption – Metrics, Methods, and Motivation, Antone KomSalesforce Adoption – Metrics, Methods, and Motivation, Antone Kom
Salesforce Adoption – Metrics, Methods, and Motivation, Antone KomCzechDreamin
 
Breaking Down the Flutterwave Scandal What You Need to Know.pdf
Breaking Down the Flutterwave Scandal What You Need to Know.pdfBreaking Down the Flutterwave Scandal What You Need to Know.pdf
Breaking Down the Flutterwave Scandal What You Need to Know.pdfUK Journal
 
Demystifying gRPC in .Net by John Staveley
Demystifying gRPC in .Net by John StaveleyDemystifying gRPC in .Net by John Staveley
Demystifying gRPC in .Net by John StaveleyJohn Staveley
 
Simplified FDO Manufacturing Flow with TPMs _ Liam at Infineon.pdf
Simplified FDO Manufacturing Flow with TPMs _ Liam at Infineon.pdfSimplified FDO Manufacturing Flow with TPMs _ Liam at Infineon.pdf
Simplified FDO Manufacturing Flow with TPMs _ Liam at Infineon.pdfFIDO Alliance
 
Easier, Faster, and More Powerful – Notes Document Properties Reimagined
Easier, Faster, and More Powerful – Notes Document Properties ReimaginedEasier, Faster, and More Powerful – Notes Document Properties Reimagined
Easier, Faster, and More Powerful – Notes Document Properties Reimaginedpanagenda
 
AI presentation and introduction - Retrieval Augmented Generation RAG 101
AI presentation and introduction - Retrieval Augmented Generation RAG 101AI presentation and introduction - Retrieval Augmented Generation RAG 101
AI presentation and introduction - Retrieval Augmented Generation RAG 101vincent683379
 
Extensible Python: Robustness through Addition - PyCon 2024
Extensible Python: Robustness through Addition - PyCon 2024Extensible Python: Robustness through Addition - PyCon 2024
Extensible Python: Robustness through Addition - PyCon 2024Patrick Viafore
 
WSO2CONMay2024OpenSourceConferenceDebrief.pptx
WSO2CONMay2024OpenSourceConferenceDebrief.pptxWSO2CONMay2024OpenSourceConferenceDebrief.pptx
WSO2CONMay2024OpenSourceConferenceDebrief.pptxJennifer Lim
 
WebAssembly is Key to Better LLM Performance
WebAssembly is Key to Better LLM PerformanceWebAssembly is Key to Better LLM Performance
WebAssembly is Key to Better LLM PerformanceSamy Fodil
 
Choosing the Right FDO Deployment Model for Your Application _ Geoffrey at In...
Choosing the Right FDO Deployment Model for Your Application _ Geoffrey at In...Choosing the Right FDO Deployment Model for Your Application _ Geoffrey at In...
Choosing the Right FDO Deployment Model for Your Application _ Geoffrey at In...FIDO Alliance
 
Secure Zero Touch enabled Edge compute with Dell NativeEdge via FDO _ Brad at...
Secure Zero Touch enabled Edge compute with Dell NativeEdge via FDO _ Brad at...Secure Zero Touch enabled Edge compute with Dell NativeEdge via FDO _ Brad at...
Secure Zero Touch enabled Edge compute with Dell NativeEdge via FDO _ Brad at...FIDO Alliance
 
Speed Wins: From Kafka to APIs in Minutes
Speed Wins: From Kafka to APIs in MinutesSpeed Wins: From Kafka to APIs in Minutes
Speed Wins: From Kafka to APIs in Minutesconfluent
 
Intro in Product Management - Коротко про професію продакт менеджера
Intro in Product Management - Коротко про професію продакт менеджераIntro in Product Management - Коротко про професію продакт менеджера
Intro in Product Management - Коротко про професію продакт менеджераMark Opanasiuk
 
Enterprise Knowledge Graphs - Data Summit 2024
Enterprise Knowledge Graphs - Data Summit 2024Enterprise Knowledge Graphs - Data Summit 2024
Enterprise Knowledge Graphs - Data Summit 2024Enterprise Knowledge
 

Último (20)

Easier, Faster, and More Powerful – Alles Neu macht der Mai -Wir durchleuchte...
Easier, Faster, and More Powerful – Alles Neu macht der Mai -Wir durchleuchte...Easier, Faster, and More Powerful – Alles Neu macht der Mai -Wir durchleuchte...
Easier, Faster, and More Powerful – Alles Neu macht der Mai -Wir durchleuchte...
 
IESVE for Early Stage Design and Planning
IESVE for Early Stage Design and PlanningIESVE for Early Stage Design and Planning
IESVE for Early Stage Design and Planning
 
ECS 2024 Teams Premium - Pretty Secure
ECS 2024   Teams Premium - Pretty SecureECS 2024   Teams Premium - Pretty Secure
ECS 2024 Teams Premium - Pretty Secure
 
Linux Foundation Edge _ Overview of FDO Software Components _ Randy at Intel.pdf
Linux Foundation Edge _ Overview of FDO Software Components _ Randy at Intel.pdfLinux Foundation Edge _ Overview of FDO Software Components _ Randy at Intel.pdf
Linux Foundation Edge _ Overview of FDO Software Components _ Randy at Intel.pdf
 
The Metaverse: Are We There Yet?
The  Metaverse:    Are   We  There  Yet?The  Metaverse:    Are   We  There  Yet?
The Metaverse: Are We There Yet?
 
Overview of Hyperledger Foundation
Overview of Hyperledger FoundationOverview of Hyperledger Foundation
Overview of Hyperledger Foundation
 
Salesforce Adoption – Metrics, Methods, and Motivation, Antone Kom
Salesforce Adoption – Metrics, Methods, and Motivation, Antone KomSalesforce Adoption – Metrics, Methods, and Motivation, Antone Kom
Salesforce Adoption – Metrics, Methods, and Motivation, Antone Kom
 
Breaking Down the Flutterwave Scandal What You Need to Know.pdf
Breaking Down the Flutterwave Scandal What You Need to Know.pdfBreaking Down the Flutterwave Scandal What You Need to Know.pdf
Breaking Down the Flutterwave Scandal What You Need to Know.pdf
 
Demystifying gRPC in .Net by John Staveley
Demystifying gRPC in .Net by John StaveleyDemystifying gRPC in .Net by John Staveley
Demystifying gRPC in .Net by John Staveley
 
Simplified FDO Manufacturing Flow with TPMs _ Liam at Infineon.pdf
Simplified FDO Manufacturing Flow with TPMs _ Liam at Infineon.pdfSimplified FDO Manufacturing Flow with TPMs _ Liam at Infineon.pdf
Simplified FDO Manufacturing Flow with TPMs _ Liam at Infineon.pdf
 
Easier, Faster, and More Powerful – Notes Document Properties Reimagined
Easier, Faster, and More Powerful – Notes Document Properties ReimaginedEasier, Faster, and More Powerful – Notes Document Properties Reimagined
Easier, Faster, and More Powerful – Notes Document Properties Reimagined
 
AI presentation and introduction - Retrieval Augmented Generation RAG 101
AI presentation and introduction - Retrieval Augmented Generation RAG 101AI presentation and introduction - Retrieval Augmented Generation RAG 101
AI presentation and introduction - Retrieval Augmented Generation RAG 101
 
Extensible Python: Robustness through Addition - PyCon 2024
Extensible Python: Robustness through Addition - PyCon 2024Extensible Python: Robustness through Addition - PyCon 2024
Extensible Python: Robustness through Addition - PyCon 2024
 
WSO2CONMay2024OpenSourceConferenceDebrief.pptx
WSO2CONMay2024OpenSourceConferenceDebrief.pptxWSO2CONMay2024OpenSourceConferenceDebrief.pptx
WSO2CONMay2024OpenSourceConferenceDebrief.pptx
 
WebAssembly is Key to Better LLM Performance
WebAssembly is Key to Better LLM PerformanceWebAssembly is Key to Better LLM Performance
WebAssembly is Key to Better LLM Performance
 
Choosing the Right FDO Deployment Model for Your Application _ Geoffrey at In...
Choosing the Right FDO Deployment Model for Your Application _ Geoffrey at In...Choosing the Right FDO Deployment Model for Your Application _ Geoffrey at In...
Choosing the Right FDO Deployment Model for Your Application _ Geoffrey at In...
 
Secure Zero Touch enabled Edge compute with Dell NativeEdge via FDO _ Brad at...
Secure Zero Touch enabled Edge compute with Dell NativeEdge via FDO _ Brad at...Secure Zero Touch enabled Edge compute with Dell NativeEdge via FDO _ Brad at...
Secure Zero Touch enabled Edge compute with Dell NativeEdge via FDO _ Brad at...
 
Speed Wins: From Kafka to APIs in Minutes
Speed Wins: From Kafka to APIs in MinutesSpeed Wins: From Kafka to APIs in Minutes
Speed Wins: From Kafka to APIs in Minutes
 
Intro in Product Management - Коротко про професію продакт менеджера
Intro in Product Management - Коротко про професію продакт менеджераIntro in Product Management - Коротко про професію продакт менеджера
Intro in Product Management - Коротко про професію продакт менеджера
 
Enterprise Knowledge Graphs - Data Summit 2024
Enterprise Knowledge Graphs - Data Summit 2024Enterprise Knowledge Graphs - Data Summit 2024
Enterprise Knowledge Graphs - Data Summit 2024
 

Fast track foundations getting serious about sequential io

  • 1. 1© Copyright 2009 EMC Corporation. All rights reserved. Fast Track Foundations: Serious about Sequential I/O James Rowland-Jones & Allan Mitchell
  • 2. 2© Copyright 2009 EMC Corporation. All rights reserved. WTF IS FAST TRACK? Introduction
  • 3. 3© Copyright 2009 EMC Corporation. All rights reserved. Blend of Hardware & Software
  • 4. 4© Copyright 2009 EMC Corporation. All rights reserved. Predictable Performance Balanced Configurations – CPU – Memory – Network – Storage
  • 5. 5© Copyright 2009 EMC Corporation. All rights reserved. Targeted to the Data Warehouse Workload Sequential Scanning of Data Leads to fewer disks
  • 6. 6© Copyright 2009 EMC Corporation. All rights reserved. Prescriptive Best Practices Recommendations You don’t get to pick and choose!
  • 7. 7© Copyright 2009 EMC Corporation. All rights reserved. WHY SQL 2008 ONLY?
  • 8. 8© Copyright 2009 EMC Corporation. All rights reserved. SQL Server 2008 Enterprise... ...was the Data Warehousing release... Compression Star Joins Few Outer Rows Parallelism Partitioned Table Parallelism Resource Governor Improved Read Ahead
  • 9. 9© Copyright 2009 EMC Corporation. All rights reserved. SQL 2008 Settings Startup Parameters – T1117 – E Resource Governor – Reduce Max memory per session Page Compression ON!
  • 10. 10© Copyright 2009 EMC Corporation. All rights reserved. Windows 2008 Settings Multipath I/O Feature is required Lock Pages In Memory Enabled
  • 11. 11© Copyright 2009 EMC Corporation. All rights reserved. SEQUENTIAL VS RANDOM I/O DEMO TIME:
  • 12. 12© Copyright 2009 EMC Corporation. All rights reserved. Counters Performance Counter  Avg. Disk Bytes / Read  Avg. Disk Sec / Read  Disk Read Bytes / Sec  Disk Reads / Sec Meaning  Size of I/Os  Latency “Seek Time”  Throughput – MB/s  I/Os per second  sys.dm_db_index_physical_file_stats  sys.dm_io_virtual_file_stats  sys.dm_io_pending_io_requests DMVs
  • 13. 13© Copyright 2009 EMC Corporation. All rights reserved. Time to Take the Red Pill...
  • 14. 14© Copyright 2009 EMC Corporation. All rights reserved.
  • 15. 15© Copyright 2009 EMC Corporation. All rights reserved. DB DESIGN CONSIDERATIONS Fast Track Thinking
  • 16. 16© Copyright 2009 EMC Corporation. All rights reserved. Indexing Think Index Light  NC Indexes aimed at low volume row lookups  Indexes are Expensive to maintain
  • 17. 17© Copyright 2009 EMC Corporation. All rights reserved. Aggregation Think Aggregation Light  Introduces Updates (fragmentation)  Removing aggs reduces maintenance
  • 18. 18© Copyright 2009 EMC Corporation. All rights reserved. Partitioning Partition Qualification Data Management P1 P4P3P2P 2
  • 19. 19© Copyright 2009 EMC Corporation. All rights reserved. RG1 RG2 RG3 RG4 RG5 RG6 RG7 RG8 RG9 RG10 Conceptual View of Fast Track Hardware 1 3 2 4 5 7 6 8 H H 4GB 4GB 4GB 4GB 4GB 4GB 4GB RAID 1 LUN 01 LUN 02 LUN 03 LUN 04 LUN 17 (Logs) LUN 05 LUN 07 LUN 06 LUN 08 SPARE SPASPB FC1 FC2 FC3 FC4 FibreSwitch LUN 09 LUN 11 LUN 10 LUN 12 LUN 18 (Logs) LUN 13 LUN 15 LUN 14 LUN 16 SPARE SPCSPD FC5 FC6 FC7 FC8 8 Core 32 GB RAM Server 4GB
  • 20. 20© Copyright 2009 EMC Corporation. All rights reserved. RG1 RG2 RG3 RG4 RG5 RG6 RG7 RG8 RG9 RG10 Performance View of Fast Track Hardware 1 3 2 4 5 7 6 8 H H LUN 01 LUN 02 LUN 03 LUN 04 LUN 17 (Logs) LUN 05 LUN 07 LUN 06 LUN 08 SPASPB FC1 FC2 FC3 FC4 FibreSwitch LUN 09 LUN 11 LUN 10 LUN 12 LUN 18 (Logs) LUN 13 LUN 15 LUN 14 LUN 16 SPCSPD FC5 FC6 FC7 FC8 200 mb/s per CPU core is central to Fast Track Architecture 32GB RAM800mb/s 800mb/s 400mb/s per port 1 LUN = 125 MB/S SP Port 650MB/s FCPort 400MB/S
  • 21. 21© Copyright 2009 EMC Corporation. All rights reserved. Performance Summary Component Throughput 8 CPU Cores 1600 MB/S 4 HBA Ports (2 adapters) 1600 MB/S 8 FC Ports (4 per array) 3200 MB/S 4 Storage Processors (2 per array) 2600 MB/S 16 LUNS (Data Only) 2000 MB/S
  • 22. 22© Copyright 2009 EMC Corporation. All rights reserved. Capacity Summary 16x 300GB Data Disks Raw (formatted) 4352 GB 4x 300GB Log Disks Raw (formatted) 1088 GB LESS RAID 1 8x RAID 1 Raid Groups for Data 2176 GB 2x RAID 1 Raid Groups for Logs 544 GB LESS TEMPDB SPACE Tempdb Data files 576 GB User Data Uncompressed Available 1600 GB Data Compression Ratio 2.5:1 4000 GB
  • 23. 23© Copyright 2009 EMC Corporation. All rights reserved. Database Layout Fi1 Fi2 Fi3 Fi4 Fi5 Fi6 Fi7 Fi9 Fi10 Fi11 Fi12 Fi13 Fi14 Fi15 Fi1 Fi2 Fi3 Fi4 Fi5 Fi6 Fi7 Fi1 Fi2 Fi3 Fi4 Fi5 Fi6 Fi7 Fi9 Fi10 Fi11 Fi12 Fi13 Fi14 Fi15 Fi9 Fi10 Fi11 Fi12 Fi13 Fi14 Fi15 tempLog Live Log Stage Log Live FG1 Stage FG1 Tempdb Primary Fi8 Fi8 Fi8 Fi16 Fi16 Fi16 Fi0 Fi0 PrimaryPrimary
  • 24. 24© Copyright 2009 EMC Corporation. All rights reserved. MANAGING FRAGMENTATION The arch enemy of Sequential I/O
  • 25. 25© Copyright 2009 EMC Corporation. All rights reserved. Fragmentation File System Extent Index
  • 26. 26© Copyright 2009 EMC Corporation. All rights reserved. File System Fragmentation Pre-Allocate your Database files Avoid O/S de-fraggers They know not what they do...
  • 27. 27© Copyright 2009 EMC Corporation. All rights reserved. Extent Fragmentation Occurs Courtesy of Concurrent DML Updates Deletes Follow the guidelines! Otherwise you’ll be re-loading the tables....
  • 28. 28© Copyright 2009 EMC Corporation. All rights reserved. Index Fragmentation  Don’t use ALTER INDEX REORGANIZE  Do one of the following – ALTER INDEX REBUILD ... SORT_IN_TEMPDB = TRUE (MAXDOP 1) – INSERT ... SELECT into a new table
  • 29. 29© Copyright 2009 EMC Corporation. All rights reserved. File Groups Isolate Volatile Data  Create a separate Filegroup  One file per LUN to spread the I/O Isolate Small / Dimension Tables Stage Data in a Separate DB or File Group
  • 30. 30© Copyright 2009 EMC Corporation. All rights reserved. DATA LOADING FOR SEQUENTIAL I/O Over to Mitchell...
  • 31. where Allan and Darren live

Notas del editor

  1. 1
  2. SMP - Symmetric MultiProcessing Scale up design that is balanced provides
  3. T1117 Ensures even growth of all files in a file group E Extent Allocation now 64 contiguous extents per file
  4. 1 Gigabit = 128 MegaBytes
  5. 1 Gigabit = 128 MegaBytes
  6. Disk blocks per database file should be kept contiguous on the physical platter within the NTFS file system
  7. 31
  8. 1 Gigabit = 128 MegaBytes