Oracle Exadata X2-8: A Critical ReviewMike AultOracle GuruTexas Memory Systems
Michael R. AultOracle Guru Nuclear Navy 6 years
 Nuclear Chemist/Programmer 10 years
 Kennedy Western University Graduate
 Bachelors Degree Computer Science
 Certified in all Oracle Versions Since 6
 Oracle DBA, author, since 1990
 Oracle AceBooks by Michael R. Ault
Statspackanalyzer.comFree Statspack/AWR Analysis Sponsored by Texas Memory SystemsLooks for IO bottlenecks and other configuration issues.
Straightforward tuning adviceWhat Is Exadata X2-8?Is it software?Is it hardware?Is it the Borg?
Exadata as Borg?
Not Really (at least not yet…)Exadata is a combination of hardware and softwareExadata takes state of the art disk, computer and flash technology and combines it with specially designed software from Oracle.Exadata hardware can only run Oracle software
Exadata X2-8 Hardware(Looks to be the SUN X4800)
Traditional Setup
Exadata SetupEach Exadata Cell has 4-96 GB flash cards
Exadata X2-8 Storage HardwareSun x4270 M2 servers that contain dual six-core Xeon L5640 processors running at 2.26 GHz, with 24GB of memory4-96 MB flash cards for Smart flash cache used to accelerate disk readsDisks are either high performance low capacity 15K or low performance high capacity 7.5KLose 66% or more of raw space for redundancy options or formatting lossesHigh Perf: 7.2 down to 2 TB per cellLow Perf: 24 down to 7 TB per cellYou pay license on a per disk basis $10,000.00/disk, $120,000.00 per cellGet IOPS based on large number of disks (168 to get 50K IOPS)200 IOPS/DISK*168 DISK=33,600 IOPS so this is doubtfulPromises of 1,000,000 IOPS from FLASH (read-only) Full X2-8 hardware cost: $1,500,000.00 (doesn’t include software!)All prices/data taken from:Oracle Exadata and Exalogic Pricelist, July 1, 2011Oracle Technology Global Pricelist, July 1, 2011, Software Investment GuideExadata Smart Flash Cache and the Sun Oracle Database Machine, Oracle, Oct 2009
Exadata X2-8 Specific Software
Something to Note:
Exadata X2-8 Oracle SoftwareLicense is per CPUApproximately $51,500.00 per CPU for all Oracle software128 CPU * $51,500.00 = $6,592,000.00Add the Cell license cost of $1,680,000.00Total license cost: $8,272,000.00Total Hardware cost: $1,500,000.00Total cost (plus shipping, handling and installation, support): $9,772,000.00All prices taken from:Oracle Exadata and Exalogic Pricelist, July 1, 2011Oracle Technology Global Pricelist, July 1, 2011, Software Investment Guide
Exadata Specific Software Features:SmartscanGoal reduce the data volume into the SGA as will only match the result set Less load on IB and RAC interconnects+CPU Predicate Filtering processes WHERE in storage Column Projections only returns columns = sql Bloom filter Joins passes matching values to storage cells to match only requested values Storage Indexes High/Low Values @1MB disk Data Mining Scoring functions run on storage
SmartscanFirst: Traditional Processing (non-indexed select):The client issues a SELECT statement with a predicate to filter and return only rows of interest. The database kernel maps this request to the file and extents containing the table being scanned. The database kernel issues the I/O to read the blocks.   All the blocks of the table being queried are read into memory. Then SQL processing is done against the raw blocks searching for the rows that satisfy the predicate.  Lastly the rows are returned to the client.From: Oracle, A Technical Overview of the Oracle Exadata Database Machine and Exadata Storage Server, April 2011
SmartscanSmartscan:The client issues a SELECT statement with a predicate to filter and return only rows of interest. The database kernel determines that Exadata storage is available and constructs an iDB command representing the SQL command issued and sends it the Exadata storage. The CELLSRV component of the Exadata software scans the data blocks to identify those rows and columns that satisfy the SQL issued. Only the rows satisfying the predicate and the requested columns are read into memory. The database kernel consolidates the result sets from across the Exadata cells. Lastly, the rows are returned to the client.From: Oracle, A Technical Overview of the Oracle Exadata Database Machine and Exadata Storage Server, April 2011
SO…Smartscan is good for full table scan issues where no indexes are availableHave heard that Oracle suggests removing indexes, but how does smartscan resolve FK-PK issues?In complex OLTP or DSS/DWH queries smartscan may not helpDepends on min/max storage indexes which can have issues with changing dataMay not be efficient on partitioned tables when a partitioned-wise join can’t be usedIn frequently updated tables smartscan may be bypassed, or, blocks will be passed that are inconsistent, requiring the DB server to filter them and retrieve the proper block from the undo segmentsFrom: Oracle: A Technical Overview of the Oracle Exadata Database Machine and Exadata Storage Server, April 2011Teradata: Exadata is Still Oracle, March 2011
Exadata Specific Software Features:Smart Flash CacheSeems to be lots of confusion between Flash cache in the server and flash cache at the DB levelFlash cache is turned on at the server in Oracle Linux and Sun using initialization parametersSmart flash cache is automatic with the Exadata cell and requires no operator intervention
Exadata Smart Flash CacheEach cell comes with 384 GB of Exadata Smart Flash Cache.In the Database Machine X2-8 and Full Rack X2-2 there is 5.3 TB of Flash Provides a 10x improvement in response time for reads and 100x improvement in IOPS for reads over regular diskDelivers a 10x blended performance average of read and write operations.Doesn’t use Least Recently Used (LRU) algorithmThe Storage Server Software keeps track of data access patterns and caches to avoid polluting the cache. Managed automatically and does not allow manual tuning. Specific objects key to performance can be pinned in cache.Only for EXADATAFrom: An Oracle White Paper, October 2009,Exadata Smart Flash Cache and the Sun Oracle Database Machine
Database Flash CacheA victim cache, it stores clean (unmodified) blocks that have been evicted from the SGA buffer cache If a block has been modified, it is first written to disk, then it is written to the Database Smart Flash Cache.Blocks can later be returned to the SGA as required. If a block cannot be found in the SGA buffer cache or the Database Smart Flash Cache, it is retrieved from disk.This is only done when DBWR has time to perform the operationsOnly works (in supported mode) on Oracle Linux and Oracle Sun SolarisSet up with two-initialization parameters:DB_FLASH_CACHE_FILEDB_FLASH_CACHE_SIZEFROM: An Oracle White Paper, September 2010, Oracle Database Smart Flash Cache
SO…Exadata flash cache can benefit both OLTP and DSS/DWHOnly for non-changing, frequently accessed dataMost helpful for readsIs basically a write-through cache using access statistics to optimize storage of hot blocksOn external flash so not as fast as database flash cache
IORM(IO Resource manager)DBRM has been enhanced for use with Exadata. DBRM lets the user define and manage intra and inter-database I/O bandwidth. Consumer groups are allocated a percent of the available I/O bandwidth and the DBRM ensures these targets are delivered. The database tags I/O with the database and consumer group.  The intra-database I/O allocations are defined and managed at the database The inter-database I/O allocations are defined within the software in the Exadata cell and managed by the I/O Resource Manager (IORM). The Exadata cell software ensures that inter-database I/O resources are managed and properly allocated within, and between, databases. Overall, DBRM ensures each database receives its specified amount of I/O resources and user defined SLAs are met.Note that IORM currently provides no management of objects cached in FlashFROM:An Oracle White Paper, April 2011, A Technical Overview of the Oracle Exadata Database Machine andExadata Storage ServerMixed Consolidated Workloads and the Sun Oracle Database Machine By Robert Stackowiak February 6, 2010
IORMDBRM are created at the DatabaseIORM are created at the CELLCreating IORM Plans on Oracle Exadata: Create a simple plan:CellCLI> alter iormplan dbplan=((name=db1_prod, level=1, allocation=70),(name=db2_prod, level=1, allocation=30),(name=db1_dev, level=2, allocation=100),(name=db1_test, level=3, allocation=70),(name=other, level=3, allocation=40)),catplan=''CellCLI> alter iormplan activeFrom: http://pauloferreiraportugal.blogspot.com/2011/04/creating-iorm-plans-on-oracle-exadata.html
Other IO FeaturesUtilizes creation of cell disks from whole physical disksCell disks are then split into ASM DISKS with HOT and COLD ASM DISKSThe Hot ASM DISKS (outer edge) are grouped together and used for hot dataThe Cold ASM DISKS are grouped together and used for cold data
SO…IORM is only for Exadata at the Cell levelIORM doesn’t regulate or understand flash cache operationsCan be used in concert with DBRM to control resources
Exadata Specific Software Features:Storage IndexesThe Storage Server Software creates and maintains the Storage Index  Tracks min and max values of columns for tables stored on that cell. When a query specifies a WHERE clause, but before any I/O, the Exadata software uses the Storage Index to determine if rows with the specified column value exist in the cell Many SQL Operations will run dramatically faster because large numbers of I/O operations are automatically replaced by a few lookups. Storage Indexes are created and maintained by the Exadata Storage Server Software.Indexes must be rebuilt every shutdown/startup operationCan cause false positives, but not false negatives
Exadata Specific Software Features:Hybrid Columnar CompressionCombination of both row and columnar methods for storing data. Achieves the compression benefits of columnar storage, avoiding the performance issues of a pure columnar format. A logical construct called the compression unit is used to store a set of Exadata Hybrid Columnar-compressed rows. When data is loaded, column values are detached from the set of rows, ordered and grouped together and then compressed. After the column data for a set of rows has been compressed, it is fit into the compression unit. Not good for volatile dataFrom: An Oracle White Paper: December 2009, Exadata Hybrid Columnar Compression
Exadata Specific Software Features:Hybrid Columnar CompressionData must be loaded using data warehouse bulk loading techniques. Exadata HCC tables can still be modified using conventional Data Manipulation Language (DML) operations, such as INSERT and UPDATE. Examples of bulk load operations commonly used in data warehouse environments are: Insert statements with the APPEND hint Parallel DML Direct Path SQL*LDR Create Table as Select (CTAS) From: An Oracle White Paper: December 2009, Exadata Hybrid Columnar Compression
Exadata Specific Software Features:HCC Warehouse CompressionWarehouse Compression Syntax CREATE TABLE emp ( emp_id NUMBER , first_name VARCHAR2(128) , last_name VARCHAR2(128) ) COMPRESS FOR QUERY [ LOW | HIGH ];Low gives 6XHigh gives 10XFrom: An Oracle White Paper: December 2009, Exadata Hybrid Columnar Compression
Exadata Specific Software Features:HCC Archive CompressionOnline Archival Compression Syntax CREATE TABLE emp ( emp_id NUMBER , first_name VARCHAR2(128) , last_name VARCHAR2(128) ) COMPRESS FOR ARCHIVE [ LOW | HIGH ];Gives up to 15XFor infrequently accessed dataWill cause notable decrease in performanceFrom: An Oracle White Paper: December 2009, Exadata Hybrid Columnar Compression
SO…HCC good for infrequently accessed dataUse OLTP compression for OLTP only 4XHave seen reports that the needed decompression/recompression due to changing data can be quite expensive performance wiseMust do initial load with pre-built blocks
Data Mining in Oracle ExadataScoring is the process of applying a data mining model to data to generate predictions. Scoring processes require significant system resources. Vast amounts of data may be involved, and algorithmic processing may be very complex.With Oracle Data Mining, scoring can be off-loaded to intelligent Oracle Exadata Storage Servers
OverallFrom: Teradata, Exadata is Still Oracle, March 2011
OverallLowSimple star schemasSimple joinsFixed dataHighHybrid or snowflake schemasComplex joins, summaries, etc.Changing data
Exadata StrengthsOne vendorHigh bandwidth data pathMoves processing (in some cases) closer to the dataOffers advanced compression for archival/fixed dataOffers 5-20X acceleration for simple queries against relatively calm dataWithin the Exadata X2-2 family, easy upgrades
Exadata WeaknessesLimited to one vendor for hardware and software (limited flexibility)Not good for rapidly changing dataExpensiveComplex hardware and softwareDaisy cutter approach to upgrade from existing system Heavy metal approach to performanceFor DP if you go Dataguard, must use Exadata
SummaryFor a new shop or project with no existing hardware Exadata is a good option if you can afford itFor consolidation Exadata may be a good fitIf you intend on keeping your licenses where they are and just getting an Exadata, it is really expensiveUses “old” technology (disks)May have issues with rapidly changing data (frequent IUD) negating the new performance featuresCan get similar or better performance improvements without getting rid of existing technologies.
Comparison of Exadata X2-8 to A RamSan710 Based Database System
Main Storage-RamSan7101-U 1-5TB SLC Flash 400,000 IOPS5 GB/s bandwidth4 QDR Infiniband ports175 microsecond latency worst case (4k)Base cost 5TB - $150,000.0012 - $1,800,000.00 (Raid 0) 30 TB
How About Some CPUs?2-SUN Fire X4800 (Same as Exadata)Each with:8-8 core 2.26 GHz  7560 CPUs1 TB memory8 PCIe modulesCost: $268,392.00
Some Licenses(Same base Oracle licenses as Exadata)128 CPUS*$51,500.00=$6,592,000No need for Cell licenses since no Cells!
Misc(Support and connection)SwitchesRackCables, etc$40,000.00
Some Database Flash cache6 – RamSan-70 Gorillas5.4 TB total of database Flash cache (2.7 TB per server)600,000 IOPS/card70 Microsecond read latency$134,400.00
TotalRamSan710s	 $1,800,000.00Servers	   $268,392.00Oracle	 $6,592,000.00RamSan70s$134,400.00Misc		    $40,000.00Total	    	 $8,834,792.00(Plus shipping, handling, installation, support)Of course this is street price, once discounts for bulk purchases and other things are removed, probably closer to 6-7 million.

Oracle Exadata X2-8: A Critical Review

  • 1.
    Oracle Exadata X2-8:A Critical ReviewMike AultOracle GuruTexas Memory Systems
  • 2.
    Michael R. AultOracleGuru Nuclear Navy 6 years
  • 3.
  • 4.
    Kennedy WesternUniversity Graduate
  • 5.
    Bachelors DegreeComputer Science
  • 6.
    Certified inall Oracle Versions Since 6
  • 7.
    Oracle DBA,author, since 1990
  • 8.
    Oracle AceBooksby Michael R. Ault
  • 9.
    Statspackanalyzer.comFree Statspack/AWR AnalysisSponsored by Texas Memory SystemsLooks for IO bottlenecks and other configuration issues.
  • 10.
    Straightforward tuning adviceWhatIs Exadata X2-8?Is it software?Is it hardware?Is it the Borg?
  • 11.
  • 12.
    Not Really (atleast not yet…)Exadata is a combination of hardware and softwareExadata takes state of the art disk, computer and flash technology and combines it with specially designed software from Oracle.Exadata hardware can only run Oracle software
  • 13.
    Exadata X2-8 Hardware(Looksto be the SUN X4800)
  • 14.
  • 15.
    Exadata SetupEach ExadataCell has 4-96 GB flash cards
  • 16.
    Exadata X2-8 StorageHardwareSun x4270 M2 servers that contain dual six-core Xeon L5640 processors running at 2.26 GHz, with 24GB of memory4-96 MB flash cards for Smart flash cache used to accelerate disk readsDisks are either high performance low capacity 15K or low performance high capacity 7.5KLose 66% or more of raw space for redundancy options or formatting lossesHigh Perf: 7.2 down to 2 TB per cellLow Perf: 24 down to 7 TB per cellYou pay license on a per disk basis $10,000.00/disk, $120,000.00 per cellGet IOPS based on large number of disks (168 to get 50K IOPS)200 IOPS/DISK*168 DISK=33,600 IOPS so this is doubtfulPromises of 1,000,000 IOPS from FLASH (read-only) Full X2-8 hardware cost: $1,500,000.00 (doesn’t include software!)All prices/data taken from:Oracle Exadata and Exalogic Pricelist, July 1, 2011Oracle Technology Global Pricelist, July 1, 2011, Software Investment GuideExadata Smart Flash Cache and the Sun Oracle Database Machine, Oracle, Oct 2009
  • 17.
  • 18.
  • 19.
    Exadata X2-8 OracleSoftwareLicense is per CPUApproximately $51,500.00 per CPU for all Oracle software128 CPU * $51,500.00 = $6,592,000.00Add the Cell license cost of $1,680,000.00Total license cost: $8,272,000.00Total Hardware cost: $1,500,000.00Total cost (plus shipping, handling and installation, support): $9,772,000.00All prices taken from:Oracle Exadata and Exalogic Pricelist, July 1, 2011Oracle Technology Global Pricelist, July 1, 2011, Software Investment Guide
  • 20.
    Exadata Specific SoftwareFeatures:SmartscanGoal reduce the data volume into the SGA as will only match the result set Less load on IB and RAC interconnects+CPU Predicate Filtering processes WHERE in storage Column Projections only returns columns = sql Bloom filter Joins passes matching values to storage cells to match only requested values Storage Indexes High/Low Values @1MB disk Data Mining Scoring functions run on storage
  • 21.
    SmartscanFirst: Traditional Processing(non-indexed select):The client issues a SELECT statement with a predicate to filter and return only rows of interest. The database kernel maps this request to the file and extents containing the table being scanned. The database kernel issues the I/O to read the blocks. All the blocks of the table being queried are read into memory. Then SQL processing is done against the raw blocks searching for the rows that satisfy the predicate. Lastly the rows are returned to the client.From: Oracle, A Technical Overview of the Oracle Exadata Database Machine and Exadata Storage Server, April 2011
  • 22.
    SmartscanSmartscan:The client issuesa SELECT statement with a predicate to filter and return only rows of interest. The database kernel determines that Exadata storage is available and constructs an iDB command representing the SQL command issued and sends it the Exadata storage. The CELLSRV component of the Exadata software scans the data blocks to identify those rows and columns that satisfy the SQL issued. Only the rows satisfying the predicate and the requested columns are read into memory. The database kernel consolidates the result sets from across the Exadata cells. Lastly, the rows are returned to the client.From: Oracle, A Technical Overview of the Oracle Exadata Database Machine and Exadata Storage Server, April 2011
  • 23.
    SO…Smartscan is goodfor full table scan issues where no indexes are availableHave heard that Oracle suggests removing indexes, but how does smartscan resolve FK-PK issues?In complex OLTP or DSS/DWH queries smartscan may not helpDepends on min/max storage indexes which can have issues with changing dataMay not be efficient on partitioned tables when a partitioned-wise join can’t be usedIn frequently updated tables smartscan may be bypassed, or, blocks will be passed that are inconsistent, requiring the DB server to filter them and retrieve the proper block from the undo segmentsFrom: Oracle: A Technical Overview of the Oracle Exadata Database Machine and Exadata Storage Server, April 2011Teradata: Exadata is Still Oracle, March 2011
  • 24.
    Exadata Specific SoftwareFeatures:Smart Flash CacheSeems to be lots of confusion between Flash cache in the server and flash cache at the DB levelFlash cache is turned on at the server in Oracle Linux and Sun using initialization parametersSmart flash cache is automatic with the Exadata cell and requires no operator intervention
  • 25.
    Exadata Smart FlashCacheEach cell comes with 384 GB of Exadata Smart Flash Cache.In the Database Machine X2-8 and Full Rack X2-2 there is 5.3 TB of Flash Provides a 10x improvement in response time for reads and 100x improvement in IOPS for reads over regular diskDelivers a 10x blended performance average of read and write operations.Doesn’t use Least Recently Used (LRU) algorithmThe Storage Server Software keeps track of data access patterns and caches to avoid polluting the cache. Managed automatically and does not allow manual tuning. Specific objects key to performance can be pinned in cache.Only for EXADATAFrom: An Oracle White Paper, October 2009,Exadata Smart Flash Cache and the Sun Oracle Database Machine
  • 26.
    Database Flash CacheAvictim cache, it stores clean (unmodified) blocks that have been evicted from the SGA buffer cache If a block has been modified, it is first written to disk, then it is written to the Database Smart Flash Cache.Blocks can later be returned to the SGA as required. If a block cannot be found in the SGA buffer cache or the Database Smart Flash Cache, it is retrieved from disk.This is only done when DBWR has time to perform the operationsOnly works (in supported mode) on Oracle Linux and Oracle Sun SolarisSet up with two-initialization parameters:DB_FLASH_CACHE_FILEDB_FLASH_CACHE_SIZEFROM: An Oracle White Paper, September 2010, Oracle Database Smart Flash Cache
  • 27.
    SO…Exadata flash cachecan benefit both OLTP and DSS/DWHOnly for non-changing, frequently accessed dataMost helpful for readsIs basically a write-through cache using access statistics to optimize storage of hot blocksOn external flash so not as fast as database flash cache
  • 28.
    IORM(IO Resource manager)DBRMhas been enhanced for use with Exadata. DBRM lets the user define and manage intra and inter-database I/O bandwidth. Consumer groups are allocated a percent of the available I/O bandwidth and the DBRM ensures these targets are delivered. The database tags I/O with the database and consumer group. The intra-database I/O allocations are defined and managed at the database The inter-database I/O allocations are defined within the software in the Exadata cell and managed by the I/O Resource Manager (IORM). The Exadata cell software ensures that inter-database I/O resources are managed and properly allocated within, and between, databases. Overall, DBRM ensures each database receives its specified amount of I/O resources and user defined SLAs are met.Note that IORM currently provides no management of objects cached in FlashFROM:An Oracle White Paper, April 2011, A Technical Overview of the Oracle Exadata Database Machine andExadata Storage ServerMixed Consolidated Workloads and the Sun Oracle Database Machine By Robert Stackowiak February 6, 2010
  • 29.
    IORMDBRM are createdat the DatabaseIORM are created at the CELLCreating IORM Plans on Oracle Exadata: Create a simple plan:CellCLI> alter iormplan dbplan=((name=db1_prod, level=1, allocation=70),(name=db2_prod, level=1, allocation=30),(name=db1_dev, level=2, allocation=100),(name=db1_test, level=3, allocation=70),(name=other, level=3, allocation=40)),catplan=''CellCLI> alter iormplan activeFrom: http://pauloferreiraportugal.blogspot.com/2011/04/creating-iorm-plans-on-oracle-exadata.html
  • 30.
    Other IO FeaturesUtilizescreation of cell disks from whole physical disksCell disks are then split into ASM DISKS with HOT and COLD ASM DISKSThe Hot ASM DISKS (outer edge) are grouped together and used for hot dataThe Cold ASM DISKS are grouped together and used for cold data
  • 31.
    SO…IORM is onlyfor Exadata at the Cell levelIORM doesn’t regulate or understand flash cache operationsCan be used in concert with DBRM to control resources
  • 32.
    Exadata Specific SoftwareFeatures:Storage IndexesThe Storage Server Software creates and maintains the Storage Index Tracks min and max values of columns for tables stored on that cell. When a query specifies a WHERE clause, but before any I/O, the Exadata software uses the Storage Index to determine if rows with the specified column value exist in the cell Many SQL Operations will run dramatically faster because large numbers of I/O operations are automatically replaced by a few lookups. Storage Indexes are created and maintained by the Exadata Storage Server Software.Indexes must be rebuilt every shutdown/startup operationCan cause false positives, but not false negatives
  • 33.
    Exadata Specific SoftwareFeatures:Hybrid Columnar CompressionCombination of both row and columnar methods for storing data. Achieves the compression benefits of columnar storage, avoiding the performance issues of a pure columnar format. A logical construct called the compression unit is used to store a set of Exadata Hybrid Columnar-compressed rows. When data is loaded, column values are detached from the set of rows, ordered and grouped together and then compressed. After the column data for a set of rows has been compressed, it is fit into the compression unit. Not good for volatile dataFrom: An Oracle White Paper: December 2009, Exadata Hybrid Columnar Compression
  • 34.
    Exadata Specific SoftwareFeatures:Hybrid Columnar CompressionData must be loaded using data warehouse bulk loading techniques. Exadata HCC tables can still be modified using conventional Data Manipulation Language (DML) operations, such as INSERT and UPDATE. Examples of bulk load operations commonly used in data warehouse environments are: Insert statements with the APPEND hint Parallel DML Direct Path SQL*LDR Create Table as Select (CTAS) From: An Oracle White Paper: December 2009, Exadata Hybrid Columnar Compression
  • 35.
    Exadata Specific SoftwareFeatures:HCC Warehouse CompressionWarehouse Compression Syntax CREATE TABLE emp ( emp_id NUMBER , first_name VARCHAR2(128) , last_name VARCHAR2(128) ) COMPRESS FOR QUERY [ LOW | HIGH ];Low gives 6XHigh gives 10XFrom: An Oracle White Paper: December 2009, Exadata Hybrid Columnar Compression
  • 36.
    Exadata Specific SoftwareFeatures:HCC Archive CompressionOnline Archival Compression Syntax CREATE TABLE emp ( emp_id NUMBER , first_name VARCHAR2(128) , last_name VARCHAR2(128) ) COMPRESS FOR ARCHIVE [ LOW | HIGH ];Gives up to 15XFor infrequently accessed dataWill cause notable decrease in performanceFrom: An Oracle White Paper: December 2009, Exadata Hybrid Columnar Compression
  • 37.
    SO…HCC good forinfrequently accessed dataUse OLTP compression for OLTP only 4XHave seen reports that the needed decompression/recompression due to changing data can be quite expensive performance wiseMust do initial load with pre-built blocks
  • 38.
    Data Mining inOracle ExadataScoring is the process of applying a data mining model to data to generate predictions. Scoring processes require significant system resources. Vast amounts of data may be involved, and algorithmic processing may be very complex.With Oracle Data Mining, scoring can be off-loaded to intelligent Oracle Exadata Storage Servers
  • 39.
    OverallFrom: Teradata, Exadatais Still Oracle, March 2011
  • 40.
    OverallLowSimple star schemasSimplejoinsFixed dataHighHybrid or snowflake schemasComplex joins, summaries, etc.Changing data
  • 41.
    Exadata StrengthsOne vendorHighbandwidth data pathMoves processing (in some cases) closer to the dataOffers advanced compression for archival/fixed dataOffers 5-20X acceleration for simple queries against relatively calm dataWithin the Exadata X2-2 family, easy upgrades
  • 42.
    Exadata WeaknessesLimited toone vendor for hardware and software (limited flexibility)Not good for rapidly changing dataExpensiveComplex hardware and softwareDaisy cutter approach to upgrade from existing system Heavy metal approach to performanceFor DP if you go Dataguard, must use Exadata
  • 43.
    SummaryFor a newshop or project with no existing hardware Exadata is a good option if you can afford itFor consolidation Exadata may be a good fitIf you intend on keeping your licenses where they are and just getting an Exadata, it is really expensiveUses “old” technology (disks)May have issues with rapidly changing data (frequent IUD) negating the new performance featuresCan get similar or better performance improvements without getting rid of existing technologies.
  • 44.
    Comparison of ExadataX2-8 to A RamSan710 Based Database System
  • 45.
    Main Storage-RamSan7101-U 1-5TBSLC Flash 400,000 IOPS5 GB/s bandwidth4 QDR Infiniband ports175 microsecond latency worst case (4k)Base cost 5TB - $150,000.0012 - $1,800,000.00 (Raid 0) 30 TB
  • 46.
    How About SomeCPUs?2-SUN Fire X4800 (Same as Exadata)Each with:8-8 core 2.26 GHz 7560 CPUs1 TB memory8 PCIe modulesCost: $268,392.00
  • 47.
    Some Licenses(Same baseOracle licenses as Exadata)128 CPUS*$51,500.00=$6,592,000No need for Cell licenses since no Cells!
  • 48.
  • 49.
    Some Database Flashcache6 – RamSan-70 Gorillas5.4 TB total of database Flash cache (2.7 TB per server)600,000 IOPS/card70 Microsecond read latency$134,400.00
  • 50.
    TotalRamSan710s $1,800,000.00Servers $268,392.00Oracle $6,592,000.00RamSan70s$134,400.00Misc $40,000.00Total $8,834,792.00(Plus shipping, handling, installation, support)Of course this is street price, once discounts for bulk purchases and other things are removed, probably closer to 6-7 million.
  • 51.
    A Million here,a Million ThereSo, what are you getting for a million less?5 GB/s*6=30 GB/s bandwidth30 TB usable raided flash with 175 microsecond latency for all transactions400,000*6=2,400,000 IOPS from Storage2,400,000 IOPS from Cache128 CPUS and 2 TB of memoryDatabase IndependenceOS independenceLess complexityBetter PerformanceLower operating costs (lower license, electric and cooling)Great performance on OLTP/DWH/DSS
  • 52.
    All-In-One-Place*Note: Various sourcesquote the Exadata X2-8 HP capacity from 28-45 usable TB. Adding 15 TB RS710 RAIDed Flash: $900,000.00
  • 53.
    SummaryExadata uses oldtechnology (disk), write through flash caches and new software to brute force performanceNew technology such as Flash and SSD as storage can get the same benefits, cheaperExadata locks you into Oracle technology and hardwareUsing newer technology keeps your options open with better performance
  • 54.
  • 55.