41
2015© Trivadis BASEL BERN BRUGG LAUSANNE ZUERICH DUESSELDORF FRANKFURT A.M. FREIBURG I.BR. HAMBURG KOPENHAGEN MUNICH STUTTGART VIENNA 2015 © Trivadis Exadata x5-2 POC with OVM: how we won against IBM Jacques Kostic Senior Consultant IMS Lausanne DOAG-19 Nov 2015 Exadata X5-2 POC with OVM 1

Poc exadata pres_doag_2015

Embed Size (px)

Citation preview

Page 1: Poc exadata pres_doag_2015

2015© Trivadis

BASEL BERN BRUGG LAUSANNE ZUERICH DUESSELDORF FRANKFURT A.M. FREIBURG I.BR. HAMBURG KOPENHAGEN MUNICH STUTTGART VIENNA

2015 © Trivadis

Exadata x5-2 POC with OVM: how we won against IBM

Jacques Kostic

Senior Consultant IMS Lausanne

DOAG-19 Nov 2015 Exadata X5-2 POC with OVM

1

Page 2: Poc exadata pres_doag_2015

2015© Trivadis

About me…

Senior Consultant, Trivadis AG, Lausanne-CH

Experience

• Oracle DBA since more than 25 years, initially with version 4

• High Availability and Backup & Recovery

• SQL and Instance Performance & Tuning

• License Audit and Consolidation

Teaching Courses at Trivadis

• Oracle Grid Infrastructure & RAC

• Oracle Data Guard

• Oracle SQL Performance & Tuning

• Oracle Instance Performance & Tuning

2 Exadata X5-2 POC with OVM Exadata X5-2 POC with OVM DOAG-19 Nov 2015

Page 3: Poc exadata pres_doag_2015

2015© Trivadis

Our company.

Exadata X5-2 POC with OVM 3

DOAG-19 Nov 2015

Trivadis is a market leader in IT consulting, system integration, solution engineering

and the provision of IT services focusing on and technologies

in Switzerland, Germany, Austria and Denmark. We offer our services in the following

strategic business fields:

Trivadis Services takes over the interacting operation of your IT systems.

O P E R A T I O N

Page 4: Poc exadata pres_doag_2015

2015© Trivadis

COPENHAGEN

MUNICH

LAUSANNE

BERN

ZURICH BRUGG

GENEVA

HAMBURG

DÜSSELDORF

FRANKFURT

STUTTGART

FREIBURG

BASEL

VIENNA

With over 600 specialists and IT experts in your region.

Exadata X5-2 POC with OVM 4

DOAG-19 Nov 2015

14 Trivadis branches and more than

600 employees

200 Service Level Agreements

Over 4,000 training participants

Research and development budget:

CHF 5.0 million

Financially self-supporting and

sustainably profitable

Experience from more than 1,900

projects per year at over 800

customers

Page 5: Poc exadata pres_doag_2015

2015© Trivadis

AGENDA

1. Introduction

2. Current Oracle Architecture

3. Alternatives with Exadata X5-2

1. Without OVM

2. With OVM

4. POC execution and results

5. Proposed architecture

6. Five years projection plan

7. Q&A

DOAG-19 Nov 2015 Exadata X5-2 POC with OVM

5

Page 6: Poc exadata pres_doag_2015

2015© Trivadis

Introduction

DOAG-19 Nov 2015 Exadata X5-2 POC with OVM

6

Page 7: Poc exadata pres_doag_2015

2015© Trivadis

Customer Overview

The name will not be disclosed but the most relevant

characteristics to the project are reported below.

Medium size customer from insurance sector

Several databases with different workload types

Lack of storage and resources with licensing constraints

Consolidation opportunities with the new Exadata X5-2

Very short time to run the POC (5 days!)

Customer

Environment

DOAG-19 Nov 2015 Exadata X5-2 POC with OVM

7

Page 8: Poc exadata pres_doag_2015

2015© Trivadis

Current Oracle architecture

DOAG-19 Nov 2015 Exadata X5-2 POC with OVM

8

Page 9: Poc exadata pres_doag_2015

2015© Trivadis

Current Oracle architecture

IBM AIX P7 PowerVM technologies, 1 LPAR per instance on uncapped CPU POOL

20 production Oracle instances

60 dev, qa, int instances

~25TB PROD/DEV/QA/INT

80 LPAR

Max 700GB per database, generally OLTP workload except for Documentum

Good SQL optimization for OLTP databases

Licensed 20 CPU Enterprise Edition with Diagnostic and Tuning packs

Uncapped CPU POOL is problematic for licensing compliance aspects

However, CPU POOL usage charts are not showing pics above 12 CPU

DOAG-19 Nov 2015 Exadata X5-2 POC with OVM

9

Page 10: Poc exadata pres_doag_2015

2015© Trivadis

Alternatives with Exadata X5-2

DOAG-19 Nov 2015 Exadata X5-2 POC with OVM

10

Page 11: Poc exadata pres_doag_2015

2015© Trivadis

Exadata X5-2: Without OVM

Pros.

Use the entire machine capacity

Less servers to manage

Pay-as-you-grow approach (COD) for software licensing is another way in which Exadata

helps to align costs with business growth

Minimum 40% of the cores must be activated

All additional options must follow the same allocation

Cons.

Isolation between databases and environments

License optimization

DOAG-19 Nov 2015 Exadata X5-2 POC with OVM

11

Page 12: Poc exadata pres_doag_2015

2015© Trivadis

Exadata X5-2: With OVM Pros.

Environment and database isolation

Hard partitioning facilitate licensing optimization

Minimum 40% of the total cores must be licensed for Enterprise Edition product

For other options, it’s linked to CPU allocation for each VM

One core per database node dedicated to dom0 (out of software licensing)

Very flexible, dynamic vCPU allocation

Allow IO resource management between all database from all virtual machines.

Db_unique_name must be unique across the entire Exadata

Cons.

Might appear more complex to manage

New feature on Exadata X5-2 (backported to X4)

Strong investment from Oracle on the technology representing the key solution for global

consolidation projects

DOAG-19 Nov 2015 Exadata X5-2 POC with OVM

12

Page 13: Poc exadata pres_doag_2015

2015© Trivadis

Exadata X5-2: With OVM

data

fra

data

fra

data

fra

free

StorageServer1 StorageServer2 StorageServer3 HD1 HD2 HD3 HD4 HD5 HD6 HD1 HD2 HD3 HD4 HD5 HD6 HD1 HD2 HD3 HD4 HD5 HD6

DBServer1 (dom0 4 vCPUs)

vClu2

vClu1

vClu3

PROD1 PROD2

PROD3 PROD5 PROD6 PROD7

QAS1 QAS2

QAS3 QAS5 QAS6

QAS7

INT1 INT2 INT3 INT5

INT6 INT7

14 vCPUs 14 vCPUs

14 vCPUs 14 vCPUs

8 vCPUs 8 vCPUs

IO Resource Manager: Category, Inter-Database, Intra-Database (db_unique_name unique across all VClusters)

PRD

QA

INT

2 Db Servers

36 cores per server

72 vCPUs per server

68 vCPUs available

DomU-1

DomU-2

DomU-3

DBServer2 (dom0 4 vCPUs)

DOAG-19 Nov 2015 Exadata X5-2 POC with OVM

13

Page 14: Poc exadata pres_doag_2015

2015© Trivadis

Exadata X5-2: OracleVM overview on Exadata

Deployment Create configuration (clusters) with Oracle Exadata Deployment Assistant (OEDA)

Configuration tool

- OEDA Configuration tool version Mar 2015 v15.084 - Patch 20645646

Prepare system

- IP allocation, customer requirements

Deploy configuration using OEDA Configuration tool

DOAG-19 Nov 2015 Exadata X5-2 POC with OVM

14

Page 15: Poc exadata pres_doag_2015

2015© Trivadis

Exadata X5-2: Cluster deployment example

DOAG-19 Nov 2015 Exadata X5-2 POC with OVM

15

Page 16: Poc exadata pres_doag_2015

2015© Trivadis

POC Execution and Result

DOAG-19 Nov 2015 Exadata X5-2 POC with OVM

16

Page 17: Poc exadata pres_doag_2015

2015© Trivadis

POC Execution and Result

Environment

Exadata 1/8

OVM Configuration

Two-node cluster with 26 vCPUs per node and 90 GB of RAM

1 database 300 GB with 30 GB of SGA (OLTP)

1 database of 700 GB with 30 GB of SGA (Documentum)

DOAG-19 Nov 2015 Exadata X5-2 POC with OVM

17

Page 18: Poc exadata pres_doag_2015

2015© Trivadis

POC Execution and Result: OLTP batch processing

Account validation batch on the OLTP database with 26 threads in parallel

*Test done with 26 vCPUs and 2 vCPUs, no differences on the execution time

Job P7 with DS8000 Exadata *Gain

Generate account validation

Preparation 2m 31s 45s 336%

Execution (28536 accounts) 2h 29m 41s 1h 17m 26s 192%

Summary generation

Execution (28536 accounts) 3h 19m 29s 2h 10m 4s 154%

Reporting

Preparation 1m 16s 48s 158%

Execution 13h 38m 45s 10h 05m 10s 135%

DOAG-19 Nov 2015 Exadata X5-2 POC with OVM

18

Page 19: Poc exadata pres_doag_2015

2015© Trivadis

POC Execution and Result: OLTP batch processing

AWR Extractions

Nothing to report!

DOAG-19 Nov 2015 Exadata X5-2 POC with OVM

19

Page 20: Poc exadata pres_doag_2015

2015© Trivadis

POC Execution and Result: Documentum

Execution time between 486 sec and 12,276 sec (average 1,226 sec)

select all doc.r_object_id, doc.a_content_type

from VFK_TST_DCTM.vfk_document_sp doc LEFT OUTER JOIN

VFK_TST_DCTM.dmi_0301d65580000206_sp ON

doc.r_object_id = dmi_0301d65580000206_sp.r_object_id

where

((doc.title!='office rendition error')

and (dmi_0301d65580000206_sp.c_status!='en traitement')

and doc.a_content_type in

('msw8', 'msw12', 'excel8book', 'excel12book', 'ppt12', 'ppt8', 'msg')

and not ( exists (select * from VFK_TST_DCTM.dmr_content_sp dmr_content

where (dmr_content.r_object_id in

(select r_object_id from VFK_TST_DCTM.dmr_content_r

where parent_id=doc.r_object_id)

and (dmr_content.full_format='pdf')

)

)

)

) and (doc.i_has_folder = 1 and doc.i_is_deleted = 0);

Query

Identify significant query

DOAG-19 Nov 2015 Exadata X5-2 POC with OVM

20

Page 21: Poc exadata pres_doag_2015

2015© Trivadis

POC Execution and Result: Documentum

Execution on production system:

----------------------------------------------------------------------------------

| Id | Operation | Name | E-Rows |

----------------------------------------------------------------------------------

| 0 | SELECT STATEMENT | | |

| 1 | NESTED LOOPS | | 499 |

| 2 | NESTED LOOPS | | 494 |

|* 3 | HASH JOIN ANTI | | 494 |

| 4 | INLIST ITERATOR | | |

|* 5 | TABLE ACCESS BY INDEX ROWID| DM_SYSOBJECT_S | 49392 |

|* 6 | INDEX RANGE SCAN | DM_SYSOBJECT_S_INDX20 | 49392 |

| 7 | VIEW | VW_SQ_1 | 31M|

| 8 | NESTED LOOPS | | 31M|

| 9 | TABLE ACCESS FULL | DMR_CONTENT_R | 69M|

|* 10 | INDEX RANGE SCAN | DMR_CONTENT_S_INDX01 | 1 |

|* 11 | INDEX UNIQUE SCAN | D_1F01D65580000924 | 1 |

|* 12 | INDEX RANGE SCAN | DMI_0301D65580000206_S_INDX06 | 1 |

----------------------------------------------------------------------------------

DOAG-19 Nov 2015 Exadata X5-2 POC with OVM

21

2h30min!

Page 22: Poc exadata pres_doag_2015

2015© Trivadis

POC Execution and Result: Documentum

Execution on Exadata:

------------------------------------------------------------------------------

| Id | Operation | Name | E-Rows |

------------------------------------------------------------------------------

| 0 | SELECT STATEMENT | | |

| 1 | NESTED LOOPS | | |

| 2 | NESTED LOOPS | | 841 |

| 3 | NESTED LOOPS | | 831 |

|* 4 | HASH JOIN ANTI | | 831 |

| 5 | INLIST ITERATOR | | |

|* 6 | TABLE ACCESS BY INDEX ROWID | DM_SYSOBJECT_S | 83083 |

|* 7 | INDEX RANGE SCAN | DM_SYSOBJECT_S_INDX20 | 83084 |

| 8 | VIEW | VW_SQ_1 | 30M|

|* 9 | HASH JOIN | | 30M|

|* 10 | INDEX STORAGE FAST FULL SCAN| DMR_CONTENT_S_INDX01 | 27M|

| 11 | TABLE ACCESS STORAGE FULL | DMR_CONTENT_R | 67M|

|* 12 | INDEX UNIQUE SCAN | D_1F01D65580000924 | 1 |

|* 13 | INDEX UNIQUE SCAN | D_1F01D65580000908 | 1 |

|* 14 | TABLE ACCESS BY INDEX ROWID | DMI_0301D65580000206_S | 1 |

------------------------------------------------------------------------------

DOAG-19 Nov 2015 Exadata X5-2 POC with OVM

22

1,21min!

Page 23: Poc exadata pres_doag_2015

2015© Trivadis

POC Execution and Result: Documentum

Different execution time

2h30 min versus 1min 21sec

Different execution plan

Missing histograms in production on column PARENT_ID for table DMR_CONTENT_R

DOAG-19 Nov 2015 Exadata X5-2 POC with OVM

23

Page 24: Poc exadata pres_doag_2015

2015© Trivadis

POC Execution and Result: Documentum

Collect missing histograms: On Exadata(just to have the elapse time)

Begin

dbms_stats.gather_table_stats (

ownname => 'VFK_TST_DCTM',

TABNAME => 'DMR_CONTENT_R',

METHOD_OPT => 'for all columns size skewonly');

End;

Elapsed: 00:01:49.905

En Prod

Begin

dbms_stats.gather_table_stats (

ownname => 'VFK_TST_DCTM',

TABNAME => 'DMR_CONTENT_R',

METHOD_OPT => 'for all columns size skewonly');

End;

Elapsed: 00:10:02.628

Factor of 5 on

the same

dataset!

DOAG-19 Nov 2015 Exadata X5-2 POC with OVM

24

Page 25: Poc exadata pres_doag_2015

2015© Trivadis

POC Execution and Result: Documentum

After having collected missing statistics, here is the result in Prod:

------------------------------------------------------------------------------

| Id | Operation | Name | E-Rows |

------------------------------------------------------------------------------

| 0 | SELECT STATEMENT | | |

| 1 | NESTED LOOPS | | |

| 2 | NESTED LOOPS | | 841 |

| 3 | NESTED LOOPS | | 831 |

|* 4 | HASH JOIN ANTI | | 831 |

| 5 | INLIST ITERATOR | | |

|* 6 | TABLE ACCESS BY INDEX ROWID | DM_SYSOBJECT_S | 83083 |

|* 7 | INDEX RANGE SCAN | DM_SYSOBJECT_S_INDX20 | 83084 |

| 8 | VIEW | VW_SQ_1 | 30M|

|* 9 | HASH JOIN | | 30M|

|* 10 | INDEX FAST FULL SCAN | DMR_CONTENT_S_INDX01 | 27M|

| 11 | TABLE ACCESS FULL | DMR_CONTENT_R | 67M|

|* 12 | INDEX UNIQUE SCAN | D_1F01D65580000924 | 1 |

|* 13 | INDEX UNIQUE SCAN | D_1F01D65580000908 | 1 |

|* 14 | TABLE ACCESS BY INDEX ROWID | DMI_0301D65580000206_S | 1 |

------------------------------------------------------------------------------

DOAG-19 Nov 2015 Exadata X5-2 POC with OVM

25

12,57min!

Page 26: Poc exadata pres_doag_2015

2015© Trivadis

POC Execution and Result: Documentum

Same execution plan on Exadata for effective comparison: optimizer_index_caching=0;

optimizer_index_cost_adj=100;

PROD 12 min 57 sec

Exadata 1 min 21 sec

Major improvement due to smart scan usage (storage clause)

------------------------------------------------------------------------------

| Id | Operation | Name | E-Rows |

------------------------------------------------------------------------------

| 0 | SELECT STATEMENT | | |

| 1 | NESTED LOOPS | | |

| 2 | NESTED LOOPS | | 841 |

| 3 | NESTED LOOPS | | 831 |

|* 4 | HASH JOIN ANTI | | 831 |

| 5 | INLIST ITERATOR | | |

|* 6 | TABLE ACCESS BY INDEX ROWID | DM_SYSOBJECT_S | 83083 |

|* 7 | INDEX RANGE SCAN | DM_SYSOBJECT_S_INDX20 | 83084 |

| 8 | VIEW | VW_SQ_1 | 30M|

|* 9 | HASH JOIN | | 30M|

|* 10 | INDEX STORAGE FAST FULL SCAN| DMR_CONTENT_S_INDX01 | 27M|

| 11 | TABLE ACCESS STORAGE FULL | DMR_CONTENT_R | 67M|

|* 12 | INDEX UNIQUE SCAN | D_1F01D65580000924 | 1 |

|* 13 | INDEX UNIQUE SCAN | D_1F01D65580000908 | 1 |

|* 14 | TABLE ACCESS BY INDEX ROWID | DMI_0301D65580000206_S | 1 |

------------------------------------------------------------------------------

Default optimizer settings

Factor 9 on the

same dataset

with the same

execution plan

DOAG-19 Nov 2015 Exadata X5-2 POC with OVM

26

Page 27: Poc exadata pres_doag_2015

2015© Trivadis

POC Execution and Result: Documentum

Change optimizer settings optimizer_index_caching=95;

optimizer_index_cost_adj=5;

PROD 4 sec

Exadata 1 sec

Less gain as smart scan is not used

------------------------------------------------------------------------------------

| Id | Operation | Name | E-Rows |

------------------------------------------------------------------------------------

| 0 | SELECT STATEMENT | | |

| 1 | NESTED LOOPS | | 4203 |

| 2 | NESTED LOOPS | | 4154 |

| 3 | INLIST ITERATOR | | |

|* 4 | TABLE ACCESS BY INDEX ROWID | DM_SYSOBJECT_S | 4154 |

|* 5 | INDEX RANGE SCAN | DM_SYSOBJECT_S_INDX20 | 4154 |

| 6 | NESTED LOOPS | | 1 |

| 7 | TABLE ACCESS BY INDEX ROWID| DMR_CONTENT_R | 2 |

|* 8 | INDEX RANGE SCAN | D_1F01D65580000005 | 2 |

|* 9 | INDEX RANGE SCAN | DMR_CONTENT_S_INDX01 | 1 |

|* 10 | INDEX UNIQUE SCAN | D_1F01D65580000924 | 1 |

|* 11 | INDEX RANGE SCAN | DMI_0301D65580000206_S_INDX04 | 1 |

------------------------------------------------------------------------------------

Factor 4 on the

same dataset

with the same

execution plan

Required by Documentum

DOAG-19 Nov 2015 Exadata X5-2 POC with OVM

27

Page 28: Poc exadata pres_doag_2015

2015© Trivadis

POC Execution and Result: Hardware tests We were requested to remove one disk! OEM Alarm

Host=sgexaadm02vm01.customer.ch

Target type=Cluster ASM

Target name=+ASM_cluster-clu1

Categories=Availability

Message=2 disks are offline.

Severity=Critical

Event report

ed time=Apr 15, 2015 10:17:05 AM CEST

Disk re-insert (rebuild) Power Power Estd Estd

INST_ID GROUP_NUMBER Operation PASS State Reqtd Actual Work Min

---------- ------------ ---------- --------- ----- ----- ------ -------- -------

2 2 REBAL RESYNC RUN 50 50 31,413 13

2 2 REBAL RESILVER WAIT 50 50 0 0

2 2 REBAL REBALANCE WAIT 50 50 0 0

2 2 REBAL COMPACT WAIT 50 50 0 0

1 2 REBAL RESYNC WAIT 50

1 2 REBAL RESILVER WAIT 50

1 2 REBAL REBALANCE WAIT 50

1 2 REBAL COMPACT WAIT 50

DOAG-19 Nov 2015 Exadata X5-2 POC with OVM

28

Page 29: Poc exadata pres_doag_2015

2015© Trivadis

POC Execution and Result: Hardware tests We were requested to unplug power cable of one storage cell! OEM Alarm Host=sgexaadm02vm01.customer.ch

Target type=Cluster ASM

Target name=+ASM_cluster-clu1

Categories=Availability

Message=Failure Group DATAC1.SGEXACELADM03 is unavailable.

Severity=Critical

Event reported time=Apr 15, 2015 5:00:25 PM CEST

Host=sgexaadm02vm01.customer.ch

Target type=Cluster ASM

Target name=+ASM_cluster-clu1

Categories=Availability

Message=Failure Group RECOC1.SGEXACELADM03 is unavailable.

Severity=Critical

Event reported time=Apr 15, 2015 5:00:25 PM CEST

Host=sgexaadm02vm01.customer.ch

Target type=Cluster ASM

Target name=+ASM_cluster-clu1

Categories=Availability

Message=12 disks are offline.

Severity=Critical

Event reported time=Apr 15, 2015 5:02:05 PM CEST

DOAG-19 Nov 2015 Exadata X5-2 POC with OVM

29

After plugging back power

cable, rebuild starts few

minutes after…

Page 30: Poc exadata pres_doag_2015

2015© Trivadis

POC Execution and Result: Conclusions

OLTP Batch

Significant gain even after huge vCPU reduction

No I/O wait events

Documentum

Major improvement when smart scan is used

Better system stability even with default optimizer settings not allays aligned with

vendor requirements

Performance increase with a factor from 4 to 9 depending if smart scan is use or not

Hardware tests

Storage protection tested and verified as requested!

DOAG-19 Nov 2015 Exadata X5-2 POC with OVM

30

Page 31: Poc exadata pres_doag_2015

2015© Trivadis

Proposed Architecture

DOAG-19 Nov 2015 Exadata X5-2 POC with OVM

31

Page 32: Poc exadata pres_doag_2015

2015© Trivadis

Exadata X5-2 based architecture Customer constraints

Isolation Secure maintenance operation

Control and adjust resource allocation

Continuity No high availability required, Data Guard protection is enough

Full capacity usage, distribute production database between the two data centers

Performance Increase performances in particular for Documentum

New application will come soon

Licensing Optimize and control licensing

Propose a five years projection plan to absorb future growth

DOAG-19 Nov 2015 Exadata X5-2 POC with OVM

32

Page 33: Poc exadata pres_doag_2015

2015© Trivadis

Exadata X5-2 based architecture with OVM

DBServer1 DBServer2

PROD1 PROD2

PROD3 PROD5 PROD6

PROD7

QAS1 QAS2

INT2

DEV1 DEV2

DRP1

IO Resource Manager: Category, Inter-Database, intra-Database (db_unique_name unique on all VClusters)

INT1

DRP1

QAS3 QAS4

INT4

DEV3 DEV4

DRP3

INT3

DRP4

10 vCPUs 10 vCPUs

4 vCPUs 4 vCPUs 4 vCPUs 4 vCPUs

2 vCPUs 2 vCPUs

vClu2

vClu1 PRD

QA/DEV vClu3

vClu4 INT

DRP

data

fra

data

fra

StorageServer1 StorageServer2 StorageServer3 HD1 HD2 HD3 HD4 HD5 HD6 HD1 HD2 HD3 HD4 HD5 HD6 HD1 HD2 HD3 HD4 HD5 HD6

data

fra

data

fra

DOAG-19 Nov 2015 Exadata X5-2 POC with OVM

33

Page 34: Poc exadata pres_doag_2015

2015© Trivadis

In total:

40 vCPUs for production databases

16 vCPUs for DEV/QA databases

16 vCPUs for INT databases

8 vCPUs for DRP databases

No additional licenses to purchase

Fix every VMs to max 14 vCPUs (to adjust power on demand)

Exadata X5-2 based architecture with OVM

Exadata X5-2 OVM Oracle infrastructure

Environment Exadata Storage Required Cores/Server Max Cores/Server Total Cores Threads CPU to License

PRD,INT,QAS,DEV,DRP I 30 TB 18 10 20 40 10

PRD,INT,QAS,DEV,DRP II 30 TB 18 10 20 40 10

Total 60 TB 25 TB 40 80 20

CPU, storage and licensing

DOAG-19 Nov 2015 Exadata X5-2 POC with OVM

34

Page 35: Poc exadata pres_doag_2015

2015© Trivadis

Dynamic host cpu reconfiguration using: xm vcpu-set

Dynamic oracle CPU_COUNT adjustment as of Oracle Oracle 12c

- Dynamic resource management update

Exadata X5-2 based architecture with OVM Adjust power on demand: MAX 14 vCPUs per VM

PROD

DBServer1

QA DEV DR

14 vCPUs

6 vCPUs

2 vCPUs

mini

DOAG-19 Nov 2015

Exadata X5-2 POC with OVM 35

Page 36: Poc exadata pres_doag_2015

2015© Trivadis

5 years projection plan

DOAG-19 Nov 2015 Exadata X5-2 POC with OVM

36

Page 37: Poc exadata pres_doag_2015

2015© Trivadis

Five Years Projection Plan

The five years projection plan is based on customer estimation with:

Up to 15% of storage increase per year

Up to 5% of processing increase per year

Solution: Exadata 1/8 de Rack with all Cores activated

DOAG-19 Nov 2015 Exadata X5-2 POC with OVM

37

Page 38: Poc exadata pres_doag_2015

2015© Trivadis

In total:

72 vCPUs for production databases

24 vCPUs for DEV/QA databases

24 vCPUs for INT databases

16 vCPUs for DRP databases

Buy 14 additional EE + Options CPU licenses to fit the needs

Fix every VMs to max 28 vCPUs (to adjust power on demand)

Five Years Projection Plan

Exadata X5-2 OVM Oracle infrastructure

Environment Exadata Storage Required Cores/Server

Max Cores/Server Total Cores Threads CPU to License

PRD,INT,QAS,DEV,DRP I 32 TB 50 TB 18 17 (18-1) 34 68 17

PRD,INT,QAS,DEV,DRP II 32 TB 50 TB 18 17 (18-1) 34 68 17

Total 64 TB 50 TB 68 136 34

CPU, storage and licensing

.

DOAG-19 Nov 2015 Exadata X5-2 POC with OVM

38

Page 39: Poc exadata pres_doag_2015

2015© Trivadis

Conclusions

The proposed solution responds perfectly to customer requirements on all areas

Environment Isolation

Performance

Capacity usage and workload adjustments

Disaster recovery

Licensing optimization

Fulfil five year projection plan requirements

DOAG-19 Nov 2015 Exadata X5-2 POC with OVM

39

Page 40: Poc exadata pres_doag_2015

2015© Trivadis

Questions...

2015 © Trivadis

BASEL BERN BRUGG LAUSANNE ZUERICH DUESSELDORF FRANKFURT A.M. FREIBURG I.BR. HAMBURG KOPENHAGEN MUNICH STUTTGART VIENNA

Jacques Kostic

Senior Consultant IMS Lausanne

DOAG-19 Nov 2015 Exadata X5-2 POC with OVM

40

Page 41: Poc exadata pres_doag_2015

2015© Trivadis

Exadata X5-2 POC with OVM 41

DOAG-19 Nov 2015

Trivadis at the DOAG 2015

Level 3 - right next to the escalator

We look forward to your visit.

Because with Trivadis you always win.