Andrejs-Running Oracle EBS I
Andrejs-Running Oracle EBS I
Journey to the
Cloud
What is New in 2022
Conference speaker:
UKOUG, nlOUG, DOAG, OATUG
UKOUG 2017 Speaker Award winner
@aprokopjevs
25 450+ 400+
Years in Business Experts across every Data Domain & Technology Global Customers
120+ Certs/Creds 125+ Certs/Accreds 60+ Certifications 70+ Certifications 50+ Certifications 20+ Certs/Accreds
6 Specializations, MSP Migration & DevOps 5 Competencies, Advisory Board Migration Factory
Badge, Partner & Competencies Incl. Data Analytics Member Certified Apps
Technical Advisory Bds and Data Platform Hosting
Infrastructure
Data Enablement Application Acceleration Business Collaboration
Modernization
Modernization and Data and analytics used to Modernized application Frictionless, secure
automation creates a drive your business, get development brings faster collaboration and sharing,
foundation for continuous better insights and better development for from anywhere on any
transformation and predictions and produce competitive advantage device drives internal
operational excellence new products transformation
Hundreds of application
100K+ VMs migrated 11,000+ databases running projects delivered Tens of thousands of users
reliably at scale migrated
Hundreds of CI/CD TCO cut by 20-60%
pipelines built Thousands of data pipelines 300k+ Google Workspace
ROI paybacks under
ERP & mission critical running seats currently managed
one year
workloads managed 500K+ new insights 350+ domains under
Applications managed on all
delivered management
Dozens of cloud platforms major clouds
Real-world machine learning
Public Cloud
Wherever your
data and Hybrid
applications reside On-Premise
ON-PREMISE
No matter what
platform you choose
CONTINUOUS
Design &
TRANSFORMATION Prove DESIGN & PROVE
Scaling through additional Focus is on collaborative design
automation and the addition of value Continuous using PaaS and existing
added services and features to Transformation automation – driven by best
address new and evolving use cases. practices.
Expand, Scale,
Manage &
Integrate
Optimize
MANAGE & OPTIMIZE
Ongoing support and continuous
transformation systems performant,
cost optimized and capable of
evolving to meet ongoing needs.
BUILD & DEPLOY
An incremental and iterative approach
Build & to developing and implementing
Deploy solutions that is driven by an agile
approach and maximizing the use of
automation.
Manage
Prof
MANAGED SERVICES Managedd
Services
Professional
Service
Services
CONSULTING
Service Deliver expert advice to solve
s
Drive ongoing performance, availability and s
the most complex challenges.
continuous systems improvement with a
Length of engagement depends
high-touch managed service model. Team- on complexity. [Fixed price or
based remote delivery with global reach hourly]
and 24x7 services available. [Open-scope or
a fixed-price]
• Database Managed Services TRANSFORMATION &
• Cloud Managed Services
• Google Cloud Managed Services IMPLEMENTATION SERVICES
Execute on a plan. (Migrate to cloud, build an
EDP, etc.) Typically team-based where
Pythian’s role can vary from full outcome
* % of Services Revenue delivery to staff augmentation. [Fixed price or
hourly]
https://www.slideshare.net/AndrejsProkopjevs/running-oracle-ebs-in-the-cloud-oaug-collaborate-18-edition
● Possible mix of IaaS, PaaS, and SaaS, depending on the platform chosen.
● All cloud platforms are similar from an infrastructure point of view (in general).
● Oracle introduced a licensing trap running its software on non-OCI cloud platforms.
● Amazon Web Services (AWS) ○ Available and not limited to platforms listed.
● Amazon Web Services (AWS) ○ Available and not limited to platforms listed.
● Features and compute capacity ● Protected private subnets (expose only what is required)
■ Secure tunnels
○ Bastion hosts
● Storage-optimized Pricing
○ Choose your own compute power parameters. ■ Link with DNS service.
The throughput limits scale together with the CPU
allocation selection. ■ OS correction start-up service (like hosts).
● Local ephemeral non-persistent storage - local disk storage with temporary data being lost during an instance restart
SQL>
● File system storage - network-based volumes (like NFS) which can be shared between multiple instances
● Object storage
The secret is in the details! Experiment and benchmark. 3. ASM with data disk rebalance.
○ Enterprise Edition
○ OCPU vs. vCPU - core vs. virtual thread
wild comparison: 1 OCPU = 2 vCPUs 2 vCPUs = 1 processor license
Important note: The above licensing outline is based on Oracle's general materials and licensing terms.
VPC features
Instance Various compute Yes Yes Yes
types type availability Flex shape availability
Performance and Depends on the instance Depends on the instance Depends on the instance
throughput limits type type type
OCPU allocation for Flex
shapes
On-demand or Universal Credits 1-year or 3-year 1-year or 3-year
reserved capacity allocation discounts allocation discounts
pricing
Performance VPUs and Dynamic Depends on the storage Depends on the storage
Performance Scaling type type
Native Oracle product PaaS and SaaS options: Custom installation and Custom installation and
integrations • Identity Cloud Service operational support operational support
• Data Integrator Cloud Service (IaaS) (IaaS)
• Analytics Cloud Service
• SOA Cloud Service
• WebCenter Portal Cloud Service
• Cloud ERP (Hybrid mode)
or
Custom installation and
operational support (IaaS)
Managed Database Service ATP-D or DBCS or ExaCS Amazon RDS No
for non-EBS Oracle (PaaS)
Oracle Database Service
for Azure (multi-cloud
integration with OCI)
Custom or third-party PaaS and SaaS options: PaaS and SaaS options: PaaS and SaaS options:
product integrations • WebLogic Cloud Service • AWS Elastic Beanstalk • Azure App Service
• Application Container Cloud Service
or or
or
Custom installation and Custom installation and
Custom installation and operational support operational support
operational support (IaaS) (IaaS) (IaaS)
• Evaluate the usage of the same-cloud platform features for heavy data exchange
integration processes to gain the best performance and avoid network latency.
○ VPUs and Dynamic Performance Scaling for storage ● EBS Cloud Manager:
○ DB-tier RAC option support with shared block storage ○ Automation scenarios may not achieve all the requirements.
Or simply do not work.
○ Exadata for the highest IO performance
● OCI Universal Credits “sold” by Oracle:
● Lift & Shift process automation tools
○ Customers will always pay at least that much.
● EBS Cloud Manager with automation options Test a lower capacity with a Pay-as-you-Go model.
○ OCI EBS Cloud Manager - advantage but far from perfect and not easily usable.
○ OS software failure - reassign DB storage volumes and restart (< 1 hour downtime).
● Automation.
○ Usually, it is the area with all the surprises and technical challenges.
○ Re-use the infrastructure and its configuration as part of the production migration.
○ 120 min - Interface switchover (with pre-built and verified targets and configuration during the pre-production
iteration). Note that this number can be even less.
● Long-term deployment:
● Automation options:
○ Automated DR failover.
● Mistakes in design will lead to technical debt and risks from day 1.
● Feature requirements
● Cost
● Licensing
○ Evaluate and understand all the options that suit your requirements.
Pythian is running its own Oracle Financials system in its on-premises data center.
Our inputs are:
● Utilization: 25 %.
Other assumptions:
● Cost estimate - current capacity and 50% less CPU/Memory power (underutilization).
Network Two regions available on the Existing presence, but only Two regions are available on the
east coast of Canada. one region in Canada for now. east coast of Canada.
Note: we are not highlighting the categories or criteria where all platform options equally apply or are not relevant.
○ But it requires additional foundational investment and training due to an existing AWS presence.
● AWS, having it as an existing cloud presence, deployment requires at least double the licensing
investment for a similar compute power capacity.
○ Due to the underutilization, the 50% capacity reduction is real to meet the current BYOL portfolio and the
budget. But it is a loss long-term.
○ It does not meet the requirements but can play an alternate DR deployment role.