Transcript CloudSIG_DatabaseCloud_V9

1
Database Consolidation Planning
and Best Practices
Nitin Vengurlekar
Charles Kim
Justin Nugent
Viscosity – CTO/Cloud Evangelist
Viscosity – CEO/Oracle Architect
Viscosity – Oracle Architect
Program Agenda
•
Private Database Cloud Business Drivers and Options
•
Private Database Cloud deployment models
•
Database Cloud Consolidation Planning and Considerations
Database Cloud Business Drivers
Lower:
• CapEx
• Servers
• Storage
•OpEx
• Maintenance
• Management
Reduce:
• Configurations
• Services
Reduce
IT Costs
Complex
ity
Standardize:
• OS
• DB Versions
Increase
Increase
Enable:
• Online changes
• Rapid response
• Faster Time to market
Reduce
Agility
Quality
of
Service
Enhance:
• IT service time
• Availability
• Security
Customers Have a Choice of Clouds
• Own & Operate, Use Managed Services, Subscribe
Private Cloud
Managed Cloud Services
Public Cloud
Build
Managed
Subscribe
Customer owns, hosts
and manages.
Customer owns.
Service provider manages.
Either may host.
Customer subscribes.
Service provider hosts
and manages.
Customers Have a Choice of Cloud Consumption
• IaaS, PaaS or SaaS
Developer
Business End User
Customizations
Customizations
Customizations
Application
Application
Platform
IaaS Cloud
SaaS Cloud
PaaS/DbaaS
Cloud
Service Provider
Service
Provider
Consumer
IT Professional
Consumer
Different Users
Private Database Cloud Architectures
Common building blocks are shared server and storage pools
OS
Hypervisor
Hypervisor
Server Consolidation
Deploy in dedicated VMs
Server virtualization
CRM
Platform
DW
Platform
Database Consolidation
Share server pool
Real Application Clusters
ERP
CRM
DB
OS
ERP
DB
DB
OS
DW
DB
ERP
DB
DW
Database Cloud
Database Cloud
DB
CRM
DB
Infrastructure Cloud
Platform
Platform
Schema/PDB Consolidation
Share database instances
Real Application Clusters
Database Consolidation Models
POLL
• How many are consolidating databases using virtualization
•
How many are consolidating databases using other
methods
•
How many of you are interested in 12c PDB
Infrastructure Cloud
Server - Provision a Database in a VM
• Reasons for adoption
• Simple to implement
• Excellent isolation
• As-is consolidation
• Legacy support
• Lower consolidation density
• Performance (latency)
• Managing sprawl
© 2014 Viscosity North America
DB
ERP
DB
•Customer concerns
CRM
DB
DW
OS
OS
OS
Hypervisor
Database Cloud
Database Consolidation – Provision Database
• Reasons for adoption
ERP
DB
DW
OS
• Customer concerns
• Requires OS standardization
• Database only
© 2011 Oracle Corporation
CRM
DB
Consolidation density
Good ROI
Performance
Supports any app
DB
•
•
•
•
OS
Private Database Cloud Deployment
Schema Consolidation - Provision a Schema/PDB
• Most efficient
• Highest consolidation density. Capability to consolidate large number of
applications.
• Efficient use of memory. Single large SGA opposed to several disparate SGAs
• Minimizes items to manage. Less OS, databases, and configurations to manage
• Reduces overall operational costs and overhead
• Extremely fast provisioning
DW
ERP
CRM
DB
• Provisioning is simply a PDB/ schema and tablespace creation
• Performance
• Native performance`
OS
OS
Private Database Cloud Deployment
Schema/PDB Consolidation Considerations
• Schema Considerations
• Requires OS and DB standardization
• Application certification required
• Isolation Management
• PDB-Multitenant Considerations
• Requires 12c Database
• Is a Cost Option
Oracle Multitenant
A new architecture for consolidating databases and simplifying operations
AP
GL
OE
Self-contained PDB for each application
• Applications run unchanged
• Rapid provisioning (via clones)
• Portability (via pluggability)
PDBs
Common operations performed at CDB
level
• Manage many as one
(upgrade, HA, backup)
• Granular control when appropriate
CDB
Root
Complementary to VMs
Shared memory and background
processes
• More applications per server
Public
13
Private Database Cloud
ROI
Greatest consolidation, maximum ROI
Database Cloud
PDB
Schema
Database
Infrastructure
Cloud
Storage
© 2011 Oracle Corporation
Servers
Consolidation
© 2011 Oracle Corporation
Cloud Architectural Assessment
Pick the architecture that best suits your needs
Business Value
Infrastructure
Consolidation
Database
Consolidation
Schema
Consolidation
PDB
Consolidation
Implementation
Easy
Easy
Difficult*
Easy
Application Suitability
Some
All
Some
Most
Isolation
Highest
High
Limited
High
Availability
High
Highest
Highest
Highest
Scalability
Limited
Excellent
Excellent
Excellent
Consolidation Density
Low
High
Highest
Highest
ROI
Low
High
Highest
Highest
* = Need to ensure application schemas can co-exist
A View of the Private Database Cloud
Cloud, An Aggregation of Cloud Pools
Cloud Pool
12c PDB Consolidation
Cloud Pool
Infrastructure Cloud
11gR1 Cloud Pool
Database Consolidation
11gR2 Cloud Pool
Schema Consolidation
11gR2 Cloud Pool
vBlock- Database Consolidation
Dev/test Cloud
Pool
Infrastructure
Cloud
DATABASE CLOUD CHANGES HOW YOU THINK
ABOUT DEPLOYMENT
Private Database Cloud Deployment
DB Cloud changes how you think about deployment
• Tenant Isolation
• Resource
• Operational
• Security
• Fault
• Bottom line : Contain “Noisy” neighbor
Private Database Cloud Deployment
DB Cloud changes how you think about deployment
•
Resource Management
– Enable Instance Caging with default resource plan
– Enable Resource Profile limits
– Enable Oracle QoS Management
• Trades-off resources just-in-time to maintain SLAs
• Ensure business critical applications get serviced and meet response
time SLAs
Private Database Cloud Deployment
DB Cloud changes how you think about deployment
• Patch and Upgrade
– Define a Patching and planned outage strategy
– Leverage Rolling Upgrades
• Minimize downtime using out of place patching
• Leverage 11.2 Standby Apply First feature
– Emergency patch scenarios for specific application
• Clone new ORACLE_HOME
• Unplug/plug schema using TTS into new database but same diskgroup
• Patch the new database and new ORACLE_HOME
Private Database Cloud Deployment
DB Cloud changes how you think about deployment
• Recovery strategies
– How you backup affects your recoverability
– Recovery Granularity Objective (RGO) becomes very important
– Recover the application using least impacting method
• Flashback technologies (Flashback transaction, table, query)
• TTS/PIT recovery or Datapump restore
• Flashback DB
Private Database Cloud Deployment
DB Cloud changes how you think about deployment
• Security Management
– Only Cloud DBA should have SYSDBA, SYSOPER and SYSASM
privileges
– Tenant DBAs should only have schema level access
– Use database roles and privileges to further limit data access
– Secure each application/schema with an Oracle Database Vault Realm
– Deploy Multi-factor rules to enforce who, where, when, and how data is
accessed
– Enable encryption on the tablespace
PLANNING AND CONSIDERATIONS
PROVISIONING MODELS AND METERING/CHARGEBACK
Provisioning
How will tenants request Cloud Services?
POLL:
• How many customers have implemented Self-Service
Provisioning?
– If so, what tools or products are you using
• How are customers implementing IT Provisioning?
– What is the workflow process
• What is the turnaround time for typically database
deployments?
Planning
Database Cloud Considerations
• What type of database provisioning
– Self Service
• Requires a robust Self Service Catalog
• Need UI for provisioning
• Provisioning Workflow
– IT Based
• Very Manual
• Need a method for queuing and reconcillation
Metering and Chargeback
Customer Consideration and Questions
POLL
• How many have or will implement chargeback
–
–
What chargeback tools are used, are they home grown or off-theshelf?
How are you measuring resource consumption?
Planning
Metering-Chargeback Considerations
• Chargeback
– Provides cost relief
– Chargeback is a good mechanism to prevent database sprawl
– Many customers have chosen not to implement any chargeback
• Show-back
–
–
–
–
Don’t actually do chargeback, but do “show-back “
Makes tenants and consumers cost conscious and consumption aware
Provides accounting of usage patterns
Can easily switch to chargeback model if needed.
Metering and Chargeback
What are your Chargeback models?
•
Typical Fixed Costs
•
•
•
•
•
Tenancy (i.e. Monthly charge, configuration)
Resource allocations (storage, CPU)
Deployment options - Data Guard, GoldenGate, etc.
Service levels
– Charge for additional service level support or higher availability
Typical Variable Costs
•
Resource utilization
– CPU Used - peak, average, time of day (common)
– Network, IOPS (rarer), Storage
Metering and Chargeback
Do you charge for operational changes?
• Service charge for operational changes
– Prevents excessive changes in the environment
– Define quantity and type of pre-paid operational requests.
• What is considered an additional request (i.e. first 3 change
requests could be built-in to the deployment)?
– Additional changes in a calendar year are subject to
cost/change
• Init.ora change
• Create test/dev environment
• Additional physical backup
• Additional Patching
PLANNING AND CONSIDERATIONS
APPLICATION READINESS AND EXECUTION
DB Cloud Planning
Application Migration Readiness
•
Applications may be in various states of technical readiness for a cloud
environment, based on application characteristics and requirements
– Name-space conflicts (relevant for schema consolidation)
– Character set requirements (relevant for schema consolidation)
– I/O rate requirements
•
Applications may have business constraints on their ability to be migrated to
the cloud
– Service level agreements
– Security
– Compliant restrictions – PCI-DSS, HIPPA, etc. datasets cannot be comingled
DB Cloud Planning
Application Migration Readiness
•
Special requirements for a given application may make it difficult to migrate
•
Decide whether the cloud will be used for new deployments only, or existing
deployments as well
•
Determine when applications will be migrated
–
–
During an existing maintenance window?
As part of a stand alone event for the application to be migrated?
•
Plan a migration process that will cause minimal disruption
•
Choose applications that will provide the biggest benefit
•
Choose “low hanging fruit” first; Go for quick wins
•
Ensure that early migrations are successful!
SUMMARY
How are Customers Deploying Private Database
Cloud
• Some call it by a different name - OaaS, DBaaS, Private
Oracle Cloud
– Many are already doing it
– However, these deployments are not formalized and “noisy neighbor”
containment is not formally managed
• Using model that best fits their licensing and skillsets.
• Cost is big factor
Summary
• Private Database Clouds consolidate servers, storage, operating systems,
databases, and workloads
– Many customers are successfully deploying Private Database Clouds today
– They are benefiting from high quality of service, low cost, reduced complexity and
increased agility
• Oracle EM 12c provides customers with end to end Cloud Management
–
–
–
–
Self-service paradigm for database deployment and management
On-demand scalability of underlying platform
Metering and chargeback for IT accountability
Extreme “agility” for developers, with “enterprise” control for IT
36