GmbH • Cloud since 2011 w/ OpenStack “Cactus” Release • Helped building many public & private cloud platforms eg. Deutsche Telekom / T-Systems ◦ Business Market Place / Open Telekom Cloud • President OpenStack DACH e.V. Twitter: @2digitsleft IRC: gpruessmann Email: [email protected] Speaker
choice • No vendor lock-in • Cloud-centric focus on success • Flexibility in technology above and below the cloud infrastructure • Rapid community innovation • Best-in-class infra, middleware and platform choices, since open cloud is all we do!
Managed On Premise Hybrid Cloud Private Cloud Public Cloud Managed Private Cloud Dedicated Off Premise Virtual On Demand Public Dedicated ± Single Tenant Platform Shared & Multi Tenant Platform
Cloud Public Cloud Private Cloud • Fully Customizable • Regulation & Total Control ◦ Data ◦ Secure by Design • Quality of Service • 24/7 Cost Efficiency • Maximum Elasticity & Scalability • Any Device - Anywhere • Low Entry Cost • Flexible Pricing/Pay as You Go • Leverage Expertise • Managed Infra & Ops • Fast Innovation • Tailored Mixing • Match Workloads to Best-Fit Infrastructure, Services & Pricing • Best In Class Services • Increase Capacity - not CapEx • Generalized Offering
2017 Source Rightscale 2017 Cloud Users running Apps in 4 (1.8/2.3) clouds and experimenting with 4 (1.8/2.1) more 9% Single Public 5% Single Private 1% No Plans 7% Multiple Private 20% Multiple Public 58% Hybrid Cloud 85% Multi Cloud Public 58% of enterprises(1000+ employees) consume hybrid cloud
by startups John Chambers: One-third of businesses today will not survive the next 10 years Transportation Watches Finance Groceries Lodging Watches Service Providers
24 6 Improved Time-to-market Reduced Cycle Time 5-6 ½-3 min Baremetal VM Improved Dev Productivity Baseline With Cloud (typical) With Cloud (max) +20-30% +60% Cloud + DevOps Accelerates the SW Dev Lifecycle > 4X Source: Actual customers & McKinsey “Reorganizing IT for Faster Software Delivery” Pre transformation Post transformation S/W iterations/ yr Months to provision infra User stories done per release
EC Alibaba Tencent .... + + Hybrid Cloud: The “Digitization” Challenge “Digitisation” Layer Apps - Value - End Customer PaaS Platform as a Service IaaS Infrastructure as a Service “internet of things” by Felix Westphal, “Autonomous Car” by Shmidt Sergey from the Noun Project 4.0 Big Data Analytics Cognitive Technology / AI Sensors & Smart Devices M2M IIoT Augmented Reality Fog Serverless …...
things” by Felix Westphal, “Autonomous Car” by Shmidt Sergey from the Noun Project 4.0 Mirantis MCP Backends / Mode 1 Private Cloud Public Cloud Fog / Edge / Device
Montoring / KPI Legal / OSS Service Registry / Discovery Support User Management Federation Governance Community Workload Transformation Service Selection & Catalog Pricing Model Identity & Access Mgmt Resource Metering Billing Cost vs. Profit Center Security API Management CSP Selection Connectivity Automation SLAs/SLOs Operational Model Integration
Montoring / KPI Legal / OSS Service Registry / Discovery Support User Management Federation Governance Community Workload Transformation Service Selection & Catalog Pricing Model Identity & Access Mgmt Resource Metering Billing Cost vs. Profit Center Security API Management CSP Selection Connectivity Automation SLAs/SLOs Operational Model Integration
Management Match ◦ Tech: Regions, Availability Zones etc. ◦ Roadmaps ◦ Hybrid: ▪ Feature Parity & Comparision of Clouds ◦ Workload Related Individual Expertise of specific CSP ▪ eg Analytics, DL, ML • Service Dependencies & Partnerships ◦ Partnerships & Sub Contractors Cloud Provider Selection Considerations 1/2 • Data Security, Data Governance and Business Policies ◦ Own Data Classification & Governance match ◦ Information Security ◦ Certification & Standards • Contracts, Pricing & SLAs ◦ Business Terms & Conditions ◦ Individual Contracts / Framework Contracts ◦ Service Delivery ◦ Legal Protections / SLAs ▪ Penalties ◦ Commercial Impact / Pricing
& Service Abstraction • Reliability & Performance ◦ Cloud Reliability Analysis ◦ Disaster Recovery • Business Health & Company Performance • Migration Support Services Cloud Provuder Selection Considerations 2/2 • Exit Planning / Sudden Death ◦ Pricing Burst / SLA meltdown ◦ Hacking ◦ Bankruptcy & Service Termination ◦ Changing Regulations & fit for service ▪ GDPR ◦ Politics “Donald Trump” by Leif Michelsen, “brexit” by David Marsh from the Noun Project Brexit “I have T-Mobile, it’s pathetic!” “The Cyber is a huge problem!”
Compute, Network, Storage, CPU, RAM … TCO vs. Agility & Time to Market TCO model considering all costs of moving to the cloud: people, tools, time, cloud migrations (automation, integration, management), training, development, security, operations, testing ... Agility Time to Market Economies of Scale Standardization Operational Efficiencies Typical Cloud Sources of “Cost Savings” “Costs” vs. “Savings” Cloud is cheap!?
un-predictable usage • Baseline Load vs. Scale Out Peaks • t2.micro vs. t2.2xlarge • Clean-Up vs. Orphans • Generic Services vs. Black Hole ◦ (Features = Convenience) • Data Gravity https://www.openstack.org/videos/boston-2017/cloud-wars-public-vs-private-cloud-economics Commercial “Cleanroom” Planning vs. Reality on Cloud
commercially reasonable efforts to make Amazon EC2 and Amazon EBS each available with a Monthly Uptime Percentage (defined below) of at least 99.95%, in each case during any monthly billing cycle (the “Service Commitment”). In the event Amazon EC2 or Amazon EBS does not meet the Service Commitment, you will be eligible to receive a Service Credit as described below. Common Cloud Services SLA Definition Source: Amazon Webservices (https://aws.amazon.com/ec2/sla/)
high level classification & acceptance further Onboarding Process Assessment Projects Requirements Cloud Readiness Analysis & Classification of Application Transformation Path/Efforts Organizational Transform. Operations & Support Model Input for Transformation & Development Project Specific Support Extended Assessment, Development & Transformation MVPs Orchestration & Automation Move & Run on Cloud Workload Onboarding Workload Onboarding Teams support projects to migrate to cloud Detect Legacy!
Trip Time 12.5 Minutes Public Cloud Server to Admin Ratio 1:2000+ Zombie Apocalypse Frequency multiple / year Tesla product upgrade cycle 11.5 days Regular “Chaos Monkey” Outage Time 0 sec
Organization Data Center (eg Virtualized) Private Cloud Virtualized infrastructure Virtualized and multi-tenancy infrastructure Process for procurement of new hardware & capacity Self-service portal Days or hours of provisioning <15 minutes provisioning time Fixed costs Pay per use or charge back Capex model Opex Model Business units take risks (underutilization) IT takes risk IT is cost center IT is profit center Operational excellence IT acts as cloud broker/sourcing integrator
Goals • Opposite MBOs • Company Goals vs. Line Goals • Non-aligned Staffing • Disparate Budgets Server Storage Diverging Goals Cloud not Prio Internal Politics
Machines Manual Approval Months Agile Continuous Commodity Hardware Automatic Workflow Days vs Solution: “Staff the Cloud team with people from the traditional teams & cloud experts - thus infecting the skilled samurai with the ninja mindset” Mode 1 Mode 2
“Products” vs. “Projects” ◦ “Product Owners” vs. “Project Managers” ◦ “Life Cycle Management” vs. “Run as built once” ◦ “Fail Fast & MVPs with Roadmaps” vs. “Traditional Waterfall / Fixed Releases” • Hybrid Cloud Product Management Process ◦ Develop Hybrid Cloud as a Product ◦ Training on Product Management, Agile Processes etc. ◦ Management Visibility & Support • DevOps vs. Siloed Operations & Support Transformation of Team, Org, Processes
◦ context, objectives and challenges of the migration ◦ Use case scenarios & application usage in the cloud ◦ coarse- and fine grained migration paths (step-by-step, component-by- component) ◦ Cloud Migration Patterns ◦ Sequence to transform the overall system architecture and the application ◦ Step by step decomposition of multi-tier application services ◦ Combination into groups of service components on the cloud ◦ integration of cloud services ◦ migration objectives ◦ Before / After Architecture Description Workload Migration & Transformation
Lift & shift, re-hosting. Cloud-Optimized Cloudification, relocation, replacement. One or more components of the application are replaced with a cloud service rather than redeveloping the architecture of the application. • Enrich application with OpenStack cloud services eg Trove • Replace specific storage services with elastic eg Swift object storage. In some cases, application may stay on the former platform but uses services from the OpenStack cloud platform. Cloud-Native Refactoring, modernization. Use the cloud to provide improved performance, scalability and elasticity to an application. A usage evaluation of components of a static or monolithic application is recommended. Complete rewrite of the application into cloud-native architecture.
GmbH Speaker • Cloud since 2011 w/ OpenStack “Cactus” Release • Helped building many public & private cloud platforms eg. Deutsche Telekom / T-Systems ◦ Business Market Place / Open Telekom Cloud • President OpenStack DACH e.V. Twitter: @2digitsleft IRC: gpruessmann Email: [email protected]
uezo
ks91
kmatsui
blue_goheimochi
yug1224
kkato1
biwakonbu
usaito
kentaitakura
yasumuusan
konifar
meow_noisy
rocio
shlominoach
robhawkes
scottboms
chrislema
davidbonilla
jacobian
mza
michaelherold
keavy
ddemaree
akosma
