Oracle Exadata v2 Fast Track Hsn ensoy Global Maksimum Data
& Information Tech Founder, VLDB Expert Global Maksimum
Data&Information Technologies 1 Not x2
Slide 2
Agenda Why do we need Exadata v2 ? Exadata Hardware Exadata
Software Better to show rather than talk. Conclusion Global
Maksimum Data&Information Technologies 2
Slide 3
3 Who am I ? Data & Information expert on VLDB environments
DWH Data Mining Inference Systems Data Archiving Solutions Niche
Storage Technologies Recovery Strategies & Solutions HA Systems
Oracle ACED on BI field Only one in Turkey Still the youngest one
all over the community. DBA of the Year 2009 7th and still the
youngest all over the world. Only one in Turkey Member of Oracle
CAB for 12g DWH development Worldwide presenter of Oracle
conferences and user group events Before completing the year HrOUG
in two weeks later Optimized Analytical Processing Capabilities of
11g Release 2 Database Consolidation Best Practices ACED Session
with Jose Seneganik Denes Kubiek UKOUG in December Optimized
Analytical Processing Capabilities of 11g Release 2
Slide 4
Global Maksimum Data&Information Technologies 4 Global
Maksimum Data & Information Technologies and Oracle Exadata v2
Only company in Turkey having IB interconnected RAC 11g
implementation experience on Linux x86-64bit. Only company in
Turkey having sufficient consultancy experience (more than 120 TB
conventional system data) on Exadata v2 Physical & Architecture
Design Migration Performance Optimization Backup & Recovery
Architectures Design Trains customers, Oracle partners, and Oracle
employees all over the Europe Strong joint relation with Oracle
Platinum Partners, Oracle Development Team Head Office, and IB
technology leaders. X X-Migrator service provider for high capacity
customers.
Slide 5
Oracle Exadata v2 Dont think Exadata as yet another product
sold by SALES guys. As a customer take it as an effortless solution
for hardware software integration. As an engineer take it as an
elegant solution of so-called unsolvable I/O problem for Oracle
databases. Global Maksimum Data&Information Technologies 5
Slide 6
Who needs Exadata v2 ? Engineers To learn that The mechanic
with a hammer thinks that all problems are nail Customers Shorter
setup time Non-Exadata Customers More stable Oracle releases Oracle
Easy to manage/standardize its code repository Global Maksimum
Data&Information Technologies 6
Slide 7
Oracle Exadata v2 Hardware Best thing about Exadata is that it
has nothing magical in it in terms of hardware. A few Sun Fire
X4170 x86-64 bit servers. A few Sun Fire X4275 x86-64 bit servers.
A few IB switches. Global Maksimum Data&Information
Technologies 7
Slide 8
Exadata v2 X-Ray Global Maksimum Data&Information
Technologies 8 42U Sun Datacenter 36-port Managed QDR IB Switched
Exadata Storage Servers Sun Fire X4170 Oracle Database Server KVM
IP Console Switch Rackmount KMM Keyboard with TFT monitor 48-port
Gigabit Ethernet Switch
Slide 9
Interconnect Network Hardware IB Switches 3 x 36-port managed
switches as opposed to Exadata v1 (2+1). 2 leaf 1 spine switches
Spine switch is only available for Full Rack because it is for
connecting multiple full racks side by side. A subnet manager
running on one switch discovers the topology of the network. HCA
Each node (RAC & Storage Cell) has a PCIe x8 40 Gbit HCA with
two ports Active-Standby Intracard Bonding. Global Maksimum
Data&Information Technologies 9
Slide 10
RAC Node Global Maksimum Data&Information Technologies 10
Sun Fire X4170 Server 2 socket Quad Core 2.53 GHz 2 Hyper-Threads
So, CPU_COUNT=16 18 DDR3 DIMM Slots 72 GB@800 MHz (2x3x3x4 GB) 4
10/100/1000Base-T Ethernet ports NET0 : Management NET1 : Public
Network NET2 : Public Network NET3 : - PCIe PES24T6G2 Switch
x8
Slide 11
Storage Node Sun Fire X4275 Server 2 socket Quad Core 2.53 GHz
6 DDR3 DIMM Slots 24 GB@1066 MHz (2x3x1x4 GB) HDD Storage 12 x
3.5-inch 600 GB 15 K RPM SAS disks 12 x 3-5-inch 2 TB 7.2 K RPM
SATA disks 4 Sun Flash Accelerator F20 PCIe Cards Global Maksimum
Data&Information Technologies 11
Slide 12
Global Maksimum Data&Information Technologies 12 Soft
Storage Node CELLSRV Multithreaded block server Buffer cache reads
Smart scans Performs I/O Resource Management Gather operational
statistics Communicates over iDB with the clients. MS OC4J
application Provides functionalities for Cell management Cell
administration Aler generation RS First process becoming live in
storage cell. Work as a hang analyzer for CELLSRV and MS MS RS
CELLSRV IORM i DB
Global Maksimum Data&Information Technologies 15 F20 PCIe
Card Not a SATA/SAS SSD driver but a x8 PCIe device providing
SATA/SAS interface. 4 Solid State Flash Disk Modules (FMod) each of
24 GB size 256 MB Cache ESM SuperCap Power Reserve
(EnergyStorageModule) provides write-back operation mode. ESM
should be enabled for optimal write performance Should be replaced
in every two years. Can be monitored using various tools like ILOM
Embedded SAS/SATA configuration will expose 16 ( 4 cards x 4 FMod )
Linux devices. /dev/sdn 4K sector boundary for Fmods Each FMod
consists of several NAND modules best performance can be reached
with multithreading (32+ thread/FMod etc)
Slide 16
Performance of F20 Global Maksimum Data&Information
Technologies 16 F20 PCIe Card (4 FMod) SequentialRead: 1.1 GB/s Max
Write: 567 MB/s (~145K IOPS @ 4K) Random @ 4KRead: 101K
IOPSWritePeak: 88K IOPSAverage : 37K IOPS Random Write Performance
Degeneration As the flash cache get full (sustained write) Wear
Leveling SLC Update Mechanism : Delete + Write Garbage Collector
write performance is degenerated due to Write Amplification. Thats
why you are not advised to put real-time performance demanding
files on flash cards Online Redo Logs
Slide 17
CapacityPerformance Global Maksimum Data&Information
Technologies 17 Aggregate Capacity Quarter Rack Half Rack Full Rack
Raw HDD SAS21 TB50 TB100 TB SATA72 TB168 TB336 TB Raw Flash1.1
TB2.6 TB5.3 TB User Data SAS6 TB14 TB28 TB SATA21 TB50 TB100 TB
Quarter Rack Half Rack Full Rack HDD Throughput SAS4.5 GB/s10.5
GB/s21GB/s SATA2.5 GB/s6 GB/s12 GB/s Flash Throughput11 GB/s25
GB/s50 GB/s Flash IOPS225,000500,0001,000,000
Slide 18
Oracle Exadata v2 Software Exadata hardware is almost
sufficient to beat any hardware configuration possible to work with
Oracle Database. But why to stop there while it is possible to do
more with Smart Scan Storage Indexes I/O Resource Manager EHCC
Global Maksimum Data&Information Technologies 18
Slide 19
Soft Components of Exadata v2 Open Soft Pieces Oracle
Enterprise Linux 5.3 Oracle defined set of RPMs Oracle OFED (bug
fixed version) Oracle Exadata Storage Software Smart Scan Encrypted
Data Data Mining Smart Flash Cache HCC Storage Index IO Resource
Manager (IORM) Oracle Exadata Bundle Patches Common Soft Pieces
Oracle RDBMS 11.2.0.1 1.Partitioning Pruning 2.Parallel Hash Join
Bloom Filtering Pairwise/Semi-pairwise Join 3.Compression HCC DBFS
Oracle Grid IS 11.2.0.1 ASM Clusterware Oracle Exadata Bundle
Patches iDB Global Maksimum Data&Information Technologies
19
Slide 20
Smart Scan Smart Scan is initially formed to be column and row
filtering based on projection and predicates. But this was just the
seed idea. Today Smart Scan can also do Projection (column)
filtering Predicate (row) filtering SELECT * FROM v$sqlfn_metadata
WHERE offloadable = 'YES'; Preperation of bloom filters for join
Smart Incremental backup Scan on encrypted data Smart File Creation
RMAN Restore Tablespace Creation File Grow Scoring for Data Mining
All data mining scoring functions are offloaded Global Maksimum
Data&Information Technologies 20
Slide 21
Smart Scan OFF. Why ? CELL_OFFLOAD_PROCESSING = FALSE The table
or partition is small. CBO doesnt choose to use direct path read.
ROW_DEPENDENCY ENABLED or rowscn is fetched. Fetch rows in rowid
order. CREATE INDEX... NOSORT LOB or LONG fetch Scan on flashback
table Cell based decryption is disable. Tablespace is not
completely on Exadata More than 255 columns are queried. Predicate
evaluation on virtual column. For dirty blocks Global Maksimum
Data&Information Technologies 21
Slide 22
Storage Index Smart Scan is about saving RAC node CPUs during
I/O processing, but storage index is about saving the processors of
Exadata storage cells. Anyhow if we figure out that T = E+W,
decreasing E in any layer will decrease T. This means faster
queries or more queries within the same period. Storage Index is
not something first used in Exadata. It is borrowed from Netezza
ZoneMap. Oracles SI is in memory It is about filtering out for a
super set of actual result set. Global Maksimum
Data&Information Technologies 22
Slide 23
First Execution Global Maksimum Data&Information
Technologies 23 select A,B,C from T1 where B