Presentation is loading. Please wait.

Presentation is loading. Please wait.

SMP MPP with PDW ** Workload requirements usually drive the architecture decision.

Similar presentations


Presentation on theme: "SMP MPP with PDW ** Workload requirements usually drive the architecture decision."— Presentation transcript:

1

2

3

4 SMP MPP with PDW ** Workload requirements usually drive the architecture decision

5

6

7

8

9 Compute Nodes Infiniband Control Nodes Passive Compute Node Dual Fiber Channel Client Drivers ETL Load Interface Corporate Backup Solution Data Center Access Corporate Network Private Network Control Rack Control Rack Data Rack Data Rack Expand by adding data rack(s) when capacity or performance requirements change Dual Active / Passive Reporting Services Analysis Services Integration Services R2 DataDirect Drivers ADO.NET, OLE-DB, ODBC DataDirect Drivers ADO.NET, OLE-DB, ODBC 1. PDW Engine 2. Admin Console 3. Metadata 4. Workspace 1. PDW Engine 2. Admin Console 3. Metadata 4. Workspace 1.User Data 2. DMS 1.User Data 2. DMS 1. SSIS Instance 2. Loader Tool 3. File Staging 1. SSIS Instance 2. Loader Tool 3. File Staging 1.File store for backups 2.Std. SQL Backups 3.Full and differential 1.File store for backups 2.Std. SQL Backups 3.Full and differential 1.Active Directory/DNS 2.HPC 3.Setup/patching 1.Active Directory/DNS 2.HPC 3.Setup/patching

10 CPU RAM HP DL360 G6 1U Intel Nehalem 8 Cores Hyper threaded 72 GB 6 – 300GB 10K SAS DELLR6101U Intel Nehalem 8 Cores Hyper threaded 96 GB 4 – 300GB 10K SAS Enterprise Class DBMS TempDB Workspace Dual Multi-Core Processors Models listed as of SQL Server 2008 R2 PDW MTP2 release ** Server models could change before RTM**

11 450 GB 15KSAS 36 TB 1 TB 7.2KSATA 80 TB DUAL 4Gb FC Stg Processor DUAL 4Gb FC Stg Processor DUAL 4Gb FC Stg Processor DUAL 4Gb FC Stg Processor Data & Log Drives (RAID 10) Data & Log Drives (RAID 10) Hot Spare Hot Spare 450 GB 15KSAS 45 TB 1 TB 7.2KSAS 100 TB Models listed as of SQL Server 2008 R2 PDW MTP2 release ** Storage models and drives could change before RTM**

12

13

14

15 15 Time Dim Date Dim ID Calendar Year Calendar Qtr Calendar Mo Calendar Day Date Dim ID Calendar Year Calendar Qtr Calendar Mo Calendar Day Store Dim Store Dim ID Store Name Store Mgr Store Size Store Dim ID Store Name Store Mgr Store Size Product Dim Prod Dim ID Prod Category Prod Sub Cat Prod Desc Prod Dim ID Prod Category Prod Sub Cat Prod Desc Sales Facts Date Dim ID Store Dim ID Prod Dim ID Mktg Camp Id Qty Sold Dollars Sold Mktg Campaign Dim Mktg Campaign Dim Mktg Camp ID Camp Name Camp Mgr Camp Start Camp End

16 16 Time Dim Date Dim ID Calendar Year Calendar Qtr Calendar Mo Calendar Day Date Dim ID Calendar Year Calendar Qtr Calendar Mo Calendar Day Store Dim Store Dim ID Store Name Store Mgr Store Size Store Dim ID Store Name Store Mgr Store Size Product Dim Prod Dim ID Prod Category Prod Sub Cat Prod Desc Prod Dim ID Prod Category Prod Sub Cat Prod Desc Sales Facts Date Dim ID Store Dim ID Prod Dim ID Mktg Camp Id Qty Sold Dollars Sold Mktg Campaign Dim Mktg Campaign Dim Mktg Camp ID Camp Name Camp Mgr Camp Start Camp End TD PD SD MDMD MDMD TD PD SD MDMD MDMD TD PD SD MDMD MDMD Smaller Dimension Tables are Replicated on Every Compute Node TD PD SD MDMD MDMD

17 17 Time Dim Date Dim ID Calendar Year Calendar Qtr Calendar Mo Calendar Day Date Dim ID Calendar Year Calendar Qtr Calendar Mo Calendar Day Store Dim Store Dim ID Store Name Store Mgr Store Size Store Dim ID Store Name Store Mgr Store Size Product Dim Prod Dim ID Prod Category Prod Sub Cat Prod Desc Prod Dim ID Prod Category Prod Sub Cat Prod Desc Sales Facts Date Dim ID Store Dim ID Prod Dim ID Mktg Camp Id Qty Sold Dollars Sold Mktg Campaign Dim Mktg Campaign Dim Mktg Camp ID Camp Name Camp Mgr Camp Start Camp End TD PD SD MDMD MDMD TD PD SD MDMD MDMD TD PD SD MDMD MDMD TD PD SD MDMD MDMD Larger Fact Table is Hash Distributed Across All Compute Nodes SF-1 SF-2 SF-3 SF-4

18 Compute Node SQL Server 2008 SP2 TempDB User Database Storage Node Storage Processor Data LUNs Tx Logs Hot Spare Distributed table hashed within node into physical tables Distributed Table SF1-a SF1-b SF1-c SF1-d SF1-e SF1-f SF1-g SF1-h SF-1 Each distribution lands on a specific LUN Replicated tables striped across LUNs SF1-a SF1-b SF1-c SF1-d SF1-e SF1-f SF1-g SF1-h Repl

19 SQL Server DW Authentication DW Configuration DW Schema DW Schema TempDB Data Movement Service (DMS) Data Movement Service (DMS) Compute Nodes Compute Node Query Tool SQL Server Data Movement Service (DMS) User Data Control Node PDW Engine Parse SQL Validate & Authorize Build MPP Plan Execute Plan Return Data to Client TempDB

20 Data Rack Control Rack 20 Control Node Landing Zone Compute Nodes Storage Nodes Infiniband Load File/SSIS Load File/SSIS DMS Ser er PDW Engine Load Manager DMS Manager DMS Manager DMS SQL Server SQL Server Load Client DMS Converter Sender Receiver Writer DMS Converter Sender Receiver Writer DWLoader invoked/ SSIS DWLoader invoked/ SSIS DMS Reads Load Data and buffers records to Send to Compute Nodes round-robin Load Manager Creates Staging Tables Each row is converted for bulk insert and hash the distribution column Hashed row is sent to appropriate node receiver for loading Received row is pushed onto writer thread Row is bulk inserted into staging table SSIS API

21 Insert-Select Load Data Bulk Insert Partitioned Staging Table (CIDX) Partitioned Staging Table (CIDX) Partitioned Final Table (CIDX) Partitioned Final Table (CIDX) Sort each BATCH in memory or TempDB Sort each BATCH in memory or TempDB Sort each partition In memory or TempDB Sort each partition In memory or TempDB Node1 Dist A Insert-Select Load Data Bulk Insert Partitioned Staging Table (CIDX) Partitioned Staging Table (CIDX) Partitioned Final Table (CIDX) Partitioned Final Table (CIDX) Sort each BATCH in memory or TempDB Sort each BATCH in memory or TempDB Sort each partition In memory or TempDB Sort each partition In memory or TempDB Node1 Dist B Insert-Select Load Data Bulk Insert Partitioned Staging Table (CIDX) Partitioned Staging Table (CIDX) Partitioned Final Table (CIDX) Partitioned Final Table (CIDX) Sort each BATCH in memory or TempDB Sort each BATCH in memory or TempDB Sort each partition In memory or TempDB Sort each partition In memory or TempDB Node1 Dist H

22

23 Credit card processing EDW Workload

24

25

26

27 www.microsoft.com/teched www.microsoft.com/learning http://microsoft.com/technet http://microsoft.com/msdn

28

29 Sign up for Tech·Ed 2011 and save $500 starting June 8 – June 31 st http://northamerica.msteched.com/registration You can also register at the North America 2011 kiosk located at registration Join us in Atlanta next year

30

31


Download ppt "SMP MPP with PDW ** Workload requirements usually drive the architecture decision."

Similar presentations


Ads by Google