23
AGILE DATA ARCHITECTURE

AGILE DATA ARCHITECTURE. Agile Data Architecture 2 Fit-for-purpose solution Enables self-service data management

Embed Size (px)

Citation preview

AGILE DATA ARCHITECTURE

Agile Data Architecture

2

Fit-for-purpose solution Enables self-service data

management

Project team A

Project team B

Primary Data Center

DR Site

Remote DC

Remote DC

Software Virtual Appliance: Flexibility, Simplicity

Deploys in minutes For any team in any location Installs on private, public

clouds

3

3 Continents, 22 Application Teams

Any HW

Any Stora

ge

Cloud Ready

Data Virtualization: Performance, Consolidation

FEATURE BENEFIT

BLOCK MAPPING

• Patented, flat metadata design scales to unlimited virtual copies

• 10-50x storage consolidation, add parallel environments at no cost

COMPRESSION • Block aware compression adds 2-4x data reduction• 2-4x reduction across virtual copies and backups

FILTERING • Intelligent filtering eliminates temporary or empty blocks• DB awareness drives 10-20% greater efficiency

CACHING• Drives performance, consolidation: 60%+ I/O load serviced

by cache• Run 20 VDBs in space of one with good performance

4

Dev: R12 Data

Mapping

QA: R12 Data

Mapping

Dev: Customization

Testing

QA: Customization Testing

1 TB

90%+ Common, Repeated DataOracle EBS Production

1 TB

1 TB

1 TB

1 TB

Unique block mapping Block aware filtering Efficient compression

90% reduction in EBS storage costs

5

.5 TB

BLOCK MAPPING COMPRESSION FILTERING

Built for VDB Performance: 20 VDBs on Average

Global “hot” read cache

Global write log SSD flash cache On-the-fly

compression, de-compression

5 sources and 35 virtual databases consolidated per Delphix server

6

CACHING

Data Orchestration: Right Data, Time, Place

FEATURE BENEFIT

SYNC • Efficiently sync heterogeneous sources in near real time • Deliver right data to right team at right time

RECORD• Synthesize, record all changes into a continuous TimeFlow• Database continuity, superior Recover Point

Objective (RPO)

PLAY

• Fast database provisioning, refresh, rollback, data integration

• Reduce time from 10 days to 10 minutes, from 4 teams to 1 team

MOVE• Promote, demote, consolidate, and recover databases• Quickly move data through application, development

lifecycle stages

REPLICATE • Efficient replication to secondary Delphix virtual appliance • High availability, disaster recovery, backup

7

Auto Sync for 10+ Heterogeneous Sources

on HP-UX EMC storage

on AIXNetApp storage

on Windows

Hitachi storage

8

on RHEL NetApp storage

10+ sources synchronize with one Delphix virtual appliance

EXADATA

Appliance

on AIXEMC storage

RAC on Solaris

NetApp storage

Remote

Any

HW

Any Stora

geCloud Ready

Remote

Cluster

on WindowsIBM storage

SYNC

Near Real Time Log Shipping

9

Local

Remote

Intra-daily refreshes of virtual copies tied to trading systems

12:36:38

12:36:38

12:36:38

12:36:41

Maintain sync within seconds of last transactions

Ensure access to fresh data Enable use of VDBs for HA

SYNC

Sync (Reduced Load)

90% load reduction with automated DB aware, change data only sync

10

5% Read I/O

5% Process I/O 5% Network Transfer Load

5% Write I/O

SYNC

1x Read I/O

1x Process I/O

1x Dump I/O

1x Read I/O 1x Transfer I/O1x Write I/O

50:1 TimeFlow (Store 50 Days in the Space of 1)

11

Quarter ending snapshots retained for 1 year

Instant recovery protects against $3000 per minute ERP downtime costs

TimeFlow

1 month retention, logs for 1 week

2 week retention, logs for 1 week

90 day full retention of logs and snapshots

Store 50 days in the space of 1 Superior SLAs (RPO and RTO) Reduce impact on systems,

networks

RECORD

Refresh, Rollback

12

on HP-UX EMC storage

Weekly scheduled refreshes, daily rollback for patch and upgrade test environments

PLAY

Aug 1, 3:36:51Provision

Current TimeRefresh

Jul 25, 6:42:18Rollback / Recover

On-demand access to source data

Superior data management SLA’s Higher application quality

Data Integration

13

6 sources integrated for MDM initiative (50% acceleration)

Aug 5, 7:28:29

Integrated Provisioning

PLAY

Integrated Rollback

Aug 1, 11:50:33

Integrated Refresh

Aug 6, 3:17:51

Decouple reliance on multiple production teams

Synchronize data delivery across sources Maximize test time and improve data quality

Promote, Demote, V2P

Test

Dev QA

14

UATProdV2P V2P

Create VDBs from VDBs V2P to restore data to physical V2P to automate UAT refresh

MOVE

Consolidate, Load Balance

Test 2Test 1

Test 3

15

30% consolidation (license reduction) of Oracle workloads

MOVE

X X

High Availability, DR, Backup for Delphix

16

X

REPLICATE

Local or Remote

Delphix as primary backup and DR solution (20+ apps)

Active-Active Multicast

Production Source

DR Site

17

Scale out to large numbers of VDBs

Enable VDB development at remote sites

REPLICATE

100+ VDBs for risk application

Self Service Enterprise Management

FEATURE BENEFIT

POLICY ENGINE

• Granular, role-based control over user, group rights management

• Easily align with enterprise policies

AUTOMATION ENGINE

• Example: init.ora file settings, scheduled refreshes• Time, labor savings and independent data access

SELF SERVICE INTERFACES

• Web GUI, CLI, web services API• Enterprise integration (branded service portals,

ticketing systems, etc.)

AUDITABILITY, SECURITY

• Comprehensive logging and reporting, preservation of source security

• Adhere to internal policies and regulatory requirements

1500 developers using VDBs for ERP projects

18

Next Steps• Identify candidate applications and projects for data

virtualization

• Complete ROI workshop with Delphix Business Solutions team– For initial target applications and full environment– Access to project team and plan needed for project impact analysis

• Delphix trial options– Cloud based

• Single Delphix virtual appliance scope– Standard onsite POC

• Multiple Delphix virtual appliance scope– Guaranteed deployment

• Production grade deployment, 90 day guarantee– Application Assessment Service

• Production grade deployment, services led adoption

19

POTENTIAL OPPORTUNITIES

ERP rollouts or projects

Agile development Application DB

upgrades Platform / DC

migrations Data warehousing /

MDM Exadata rollout Private cloud build-out Secure test data

Thank You

Performance Optimization Measures

• Overall labor and process gains• Global “Hot” Read Cache• On-the-fly Compression, De-compression

Virtualization overhead offset by:

21

w.o. Delphix

w. Delphix

I/O processingtime

Provisioning and refreshes – actual labor

Provisioning and refreshes - scheduling and process overhead

Connectivity Protocols

22

Source Server

Target Servers

SSHJDBC

SSHJDBC

HTTP

NFS

Compressed Data

SSH

FC, iSCSI

System Requirements

Delphix Server– Virtualization Platform: VMware (ESX Server 3.x, vSphere 4.x, vSphere

5.x)

– VM Guest Container Profile: Solaris 10 64-bit

– Virtual CPUs: 8 vCPUs

– Memory: 16 GB (minimum), add more if indicated

Source and Target Servers– Oracle 9.2.0.8, Oracle 10g R2, Oracle 11g R1, or Oracle 11g R2

– Solaris, RedHat, OEL, SUSE, AIX, HPUX

23