Introduction
Problem statement
Cloud computing considerations for companies framework motivation
Paper organisation
Related work
Cloud adoption frameworks comparison
Cloud Framework Evaluation Criteria/Frameworks | CCCC Cloud Adoption Framework | ARTIST Methodology and Framework [9] | Migration to PaaS clouds - Migration process and architectural concerns [17] | Cloud Calculator: A cloud assessment tool for the public administration [24] | A knowledge management-based cloud computing adoption decision making framework [14] |
---|---|---|---|---|---|
Analysing Context – identify business value and relevant stakeholders to quantify if there is business value in moving the application to cloud | Fully Supported | Fully Supported | Fully Supported | Not Supported | Fully Supported |
Understanding Legacy Applications – capture an abstract of the as-is application | Fully Supported | Fully Supported | Fully Supported | Fully Supported | Partially Supported |
Analysing Migration Requirements – Determine if a prototype is required to manage technical risks | Fully Supported | Fully Supported | Partially Supported | Partially Supported | Partially Supported |
Planning Migration – Develop a migration plan for the legacy application | Fully Supported | Partially Supported | Partially Supported | Partially Supported | Partially Supported |
Cloud Service Platform Selection – Identify the cloud platform features required by the application | Fully Supported | Partially Supported | Fully Supported | Partially Supported | Fully Supported |
Migration Type – Determine migration approach from the least intrusive: ‘lift and shirt’ to the most intrusive: cloud native transformation | Fully Supported | Partially Supported | Partially Supported | Not Supported | Not Supported |
Unit of Migration – Identify any components that are subject to regulations or compliance that are not suitable for cloud | Fully Supported | Not Supported | Not Supported | Not Supported | Not Supported |
Training – understanding new documentation, Application Programming Interfaces, tools and data movement charges | Partially Supported | Not Supported | Partially Supported | Not Supported | Partially Supported |
Re-architecting Legacy Applications – Specify the Target architecture model that addresses transformation requirements | Fully Supported | Partially Supported | Partially Supported | Partially Supported | Not Supported |
Testing and Continuous Integration – test adaptions from a functional and non-functional perspective as they are made | Not Supported | Not Supported | Not Supported | Not Supported | Not Supported |
Enterprise strategy, cloud migration and cloud economics
Using cloud standards to avoid lock-in
Cloud SLAs and SLA metrics
CCCC framework
Cloud decision framework
Cloud computing considerations for companies (CCCC) framework overview
-
Application silos per business unit with a nest of interconnections are discovered that require consideration to identify the dependencies and therefore allocate applications into migration waves to ensure end users are not impacted during the migration. Waves are essentially ‘sub-projects’ or self-contained work tasks with virtual or physical server to application grouping [44].
-
Fragile and brittle coupling between applications such as point to point integration that have hard coded configurations that must be updated during the migration process.
-
Strategic decision making for application portfolios – by incorporation of two new processes of APP and APA to support Cloud Computing adoption across application portfolios to enable the identification of legacy applications suitable for cloud computing from a technical and economic perspective while understanding and allowing for their dependencies in forecasting migration and cloud run costs using:
-
Application Portfolio Profiling – this phase involves interviews with application stakeholders to obtain business plans for each application such as ‘retire soon’ or identify the application as strategic, followed by detailed data gathering to establish baseline information from which an APA is based to obtain TPR and FVA for each application
-
Application Portfolio Assessment – this phase involves identification of each application’s dependencies, application role, application affinity, cost saving opportunities such as duplicate licensing and consolidating managed services or service desks
-
Enterprise Cloud Computing Adoption Recommendation - expansion of the tactical streamlined decision-making to support enterprise Cloud Computing adoption:
-
Assess cloud adoption affordability, incorporating the hidden costs of cloud computing, using the output of the CCCC framework that enables comparison between public and private cloud costs for each application to provide stakeholders the information they may need for due diligence purposes
-
Assess if the savings are enough to pay for the migration and transformation of applications to Hybrid Cloud
-
The business case formula to determine if the planned savings cover the investment required to migrate and/or transform the applications to Hybrid Cloud is: As-Is Compute/Network/Storage costs compared with the projected public and private cloud IaaS costs for Year 1, Year 2, Year 3, Year 4 and Year 5. If the savings from the move to Cloud Computing in Year 1, Year 2 and Year 3 cover all or most of the Application Transformation and transition costs, then Enterprise Cloud Computing adoption is considered feasible and recommended.
Application portfolio profiling
Interviewing business stakeholders
Data collection
Application filtering
-
Software licensing agreements and operating system choice are reviewed for each application, and those applications that may not be fit for purpose or redundant in a cloud computing model are excluded from the assessment
-
For those applications that are targeted for retirement or that are no longer strategic, they are not included in the assessment
Capturing TPRs and FVAs
Application portfolio assessment
-
Validation of the current physical and virtual server environment
-
Sizing of target environment
-
Identification and establishment of application linkage/communications from an ‘as-is’ environment and ‘to-be’ environment
-
Building an application migration wave plan based on Application Affinity and defining target architecture configurations with the aim of minimizing migration risk. This involves mapping applications to a functional map to identify any duplication in functionality in the portfolio and allocating applications to an application grouping and therefore a migration wave group.
-
Identifying key opportunities for transformation to exploit the Cloud Computing platform
-
Implementing a security design, such as, encryption, zoning map, professional services partner access to the environment, procedures and processes
Framework illustration
-
Q1: Which applications in the portfolio are suitable for Cloud Computing?
-
Q2: What is the migration impact in terms of cost and duration?
-
Q3: What are the cost savings of moving suitable applications to the Cloud Computing?
-
Q4: What is the split of applications being deployed between public and private cloud?
Data collection
Application | Role | Environment | Operating System | vCPU | RAM | Quantity |
---|---|---|---|---|---|---|
SAP | ERP | Test | Windows | 4 | 14 | 1 |
SAP | ERP | Development | Windows | 4 | 14 | 1 |
SAP | E0RP | Production | Windows | 16 | 56 | 1 |
S4/Hana | Business Suite | Development | SuSE Linux | 4 | 14 | 1 |
S4/Hana | Business Suite | Test | SuSE Linux | 16 | 128 | 1 |
S4/Hana | Business Suite | Production | SuSE Linux | 32 | 256 | 1 |
SAP | Business Connector | Development | Windows | 2 | 7 | 1 |
SAP | Business Connector | Production | Windows | 2 | 7 | 1 |
SAP | Print Server | Production | Windows | 2 | 7 | 1 |
Shared | Active Directory | Production | Windows | 2 | 7 | 1 |
Shared | Active Directory | Disaster Recovery | Windows | 2 | 7 | 1 |
-
Upstream/Downstream dependencies of the applications – Active Directory
-
Complexity (Simple/Medium/Complex) & Criticality – Complex and Critical (based on interview with application owner)
-
Review of Application sunset schedule – SAP is strategic (based on interview with application owner)
-
Business Impact i.e. SLA, DR/BC – High
-
For accurate utilisation data, the application discovery is scheduled over a four to six-week period as it apparent when the peak periods of utilisation are. If discovery was not performed twice daily, then peak periods are at risk of being missed.
-
The application data is collected both during business hours and out of business hours to ensure coverage of utilisation data. If discovery is performed only during one period of the day, the peak utilisation periods could be missed.
-
The findings are validated with the stakeholders to determine if any of them were unexpected as discovered data may be exception for the time of year it is taken.
Technical platform recommendation
Scenario | SAP | |
---|---|---|
Cloud Decision Criteria | Criteria | Criteria Classification |
Availability | Required | |
Business Service Availability | Required | |
Long running business process | Required | |
Application Usage | Required | |
Regulatory requirements | Required | |
Operating Costs | Required | |
Performance | Required | |
Application architecture | Required | |
Application constraints | Required | |
Security | Required | |
Data Security Classification | Required | |
Network Global Load Balancing | Optional | |
Connectivity to private MPLS network or internet VPN | Required | |
Hypervisor | Required | |
Enterprise Control | Required | |
Data Classification | Required | |
Technology Standardisation | Required |
Private Cloud | Public Cloud | No Change | |
---|---|---|---|
Applications | 204 | 93 | 53 |
Financial viability assessment
Public Cloud Cost Items | ||
Virtual Compute | Storage | Ingress |
Egress | Backup | Audit |
DNS Services | Elasticity | Custom Items |
Private Cloud Cost Items | ||
Compute | Storage | Data Centre Space |
Air Conditioning | Audit | Power |
Backup | Installation | Software licensing |
Frame | Custom Items |
Cost Item | Public Cloud | Private Cloud | Relevance |
---|---|---|---|
Architecture & Project Management | |||
Solution Architecture | √ | √ | The work required to ensure the application, infrastructure, data and service management solution are cohesive. |
Project Management | √ | √ | Project Managers are charged with planning, budgeting and resourcing. |
Network Connectivity, Security | |||
Security Services | √ | √ | Provides consideration for new firewall, anti-virus or Denial of Service |
DNS Services | √ | √ | A scalable and highly available domain name that is part of a public or private cloud platform service |
Database as a Service | √ | √ | An example of a custom entry in the cost model |
Connectivity to a private network | X | √ | Local access from a data centre to the client’s private network |
Monitoring & Management | |||
Lifecycle Management | √ | X | It is an IT Asset lifecycle is a sequence of events to determine if the asset requires refresh or replacement |
Automation Components | √ | √ | Components used to automate a manual process or part thereof. |
Governance of resources | √ | √ | A method of providing role-based access for rights for the provisioning of cloud resources. |
Commercial Considerations | |||
PAYG Commercial Model | √ | X | A flexible payment arrangement that provides consumption on a granular basis |
Reserved Instances Commercial Model | √ | X | A dedicated resource is a reservation of resources and capacity typically over a multi-year commitment in public cloud. |
Managed Private Cloud providing a 30% resource overhead to the baseline requirement | X | √ | We have used a heuristic of 30% that balances elasticity, peak loads, professional services cost and overhead in the private cloud to account |
Elasticity of Public Cloud | √ | X | Enable periodic or seasonal requirements for elasticity in public cloud to be captured to balance this against the resource overhead in Managed Private Cloud |
High Availability | √ | √ | Public Cloud and Managed Cloud providers will have dependencies upon components to meet high availability requirements. |
Custom Items | √ | √ | Any items considered important by the Cloud Adopter to be included that are above the IaaS layer and specific to a public or private cloud. |
Up-front | Monthly | End of contract | Qty | Year | |||||||||||||||
---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
0 | 1 | 3 | 4 | 5 | 6 | ||||||||||||||
Up-front | Accumulative | Accumulative | Accumulative | Accumulative | Accumulative | ||||||||||||||
Architecture & PM | |||||||||||||||||||
Solution Architecture | $ | – | $ | – | $ | – | 0 | $ | – | $ | – | $ | – | $ | – | $ | – | $ | |
Project Management | $ | – | $ | – | $ | – | 0 | $ | – | $ | – | $ | – | $ | – | $ | – | $ | |
Connectivity | $ | – | $ | – | $ | – | 0 | $ | – | $ | – | $ | – | $ | – | $ | – | $ | |
Design & Implement (PROD) | $ | 1,295,000.00 | $ | – | $ | – | 2 | $ | 2,590,000.00 | $ | 2,590,000.00 | $ | 2,590,000.00 | $ | 2,590,000.00 | $ | 2,590,000.00 | $ | 2,590,000.00 |
Design & Implement (DR) | $ | – | $ | – | $ | – | 0 | $ | – | $ | – | $ | – | $ | – | $ | – | $ | |
Compute & Storage (PROD) | |||||||||||||||||||
Vblock 350 Frame | $ | – | $ | 29,450.00 | $ | – | 2 | $ | – | $ | 706,800.00 | $ | 1,413,600.00 | $ | 2,120,400.00 | $ | 2,827,200.00 | $ | 3,534,000.00 |
Virtual Blades | $ | – | $ | 2450.00 | $ | – | 60 | $ | – | $ | 1,764,000.00 | $ | 3,528,000.00 | $ | 5,292,000.00 | $ | 7,056,000.00 | $ | 8,820,000.00 |
Physical Blades | $ | – | $ | 1150.00 | $ | – | 0 | $ | – | $ | – | $ | – | $ | – | $ | – | $ | |
Storage | $ | – | $ | 7450.00 | $ | – | 1 | $ | – | $ | 89,400.00 | $ | 178,800.00 | $ | 268,200.00 | $ | 357,600.00 | $ | 447,000.00 |
Compute & Storage (DR) | |||||||||||||||||||
Vblock 350 Frame | $ | – | $ | 29,450.00 | $ | – | 2 | $ | – | $ | 706,800.00 | $ | 1,413,600.00 | $ | 2,120,400.00 | $ | 2,827,200.00 | $ | 3,534,000.00 |
Virtual Blades | $ | – | $ | 2450.00 | $ | – | 60 | $ | – | $ | 1,764,000.00 | $ | 3,528,000.00 | $ | 5,292,000.00 | $ | 7,056,000.00 | $ | 8,820,000.00 |
Physical Blades | $ | – | $ | 1150.00 | $ | – | 0 | $ | – | $ | – | $ | – | $ | – | $ | – | $ | |
Storage | $ | – | $ | 7450.00 | $ | – | 1 | $ | – | $ | 89,400.00 | $ | 178,800.00 | $ | 268,200.00 | $ | 357,600.00 | $ | 447,000.00 |
Ingress & Egress Fees | $ | – | $ | – | $ | – | $ | – | $ | – | $ | – | $ | – | $ | – | $ | ||
Monitoring & Management | $ | – | $ | 25,000.00 | $ | – | 19 | $ | – | $ | 5,700,000.00 | $ | 11,400,000.00 | $ | 17,100,000.00 | $ | 22,800,000.00 | $ | 28,500,000.00 |
Data centre | |||||||||||||||||||
Power | $ | – | $ | – | $ | – | 0 | $ | – | $ | – | $ | – | $ | – | $ | – | $ | |
Cooling | $ | – | $ | – | $ | – | 0 | $ | – | $ | – | $ | – | $ | – | $ | – | $ | |
Floor space | $ | – | $ | – | $ | – | 0 | $ | – | $ | – | $ | – | $ | – | $ | – | $ | |
Lifecycle Management | |||||||||||||||||||
Platform engineering | $ | 1200.00 | $ | – | 0.5 | $ | – | $ | 7200.00 | $ | 14,400.00 | $ | 21,600.00 | $ | 28,800.00 | $ | 36,000.00 | ||
Incident management & remediation | $ | – | $ | – | $ | – | 0 | $ | – | $ | – | $ | – | $ | – | $ | – | $ | |
Platform migration | $ | 1200.00 | $ | – | $ | 1200.00 20 | $ | 24,000.00 | $ | 24,000.00 | $ | 24,000.00 | $ | 24,000.00 | $ | 24,000.00 | $ | 24,000.00 | |
Predeployment validation lab | $ | – | $ | – | 1 | $ | – | $ | – | $ | – | $ | – | $ | – | $ | |||
Backup | $ | – | $ | 128.57 | $ | – | 44 | $ | – | $ | 67,885.71 | $ | 135,771.43 | $ | 203,657.14 | $ | 271,542.86 | $ | 339,428.57 |
Contract Administration | $ | – | $ | 500.00 | $ | – | 1 | $ | – | $ | 6000.00 | $ | 12,000.00 | $ | 18,000.00 | $ | 24,000.00 | $ | 30,000.00 |
Audit | $ | – | $ | 300.00 | $ | – | 1 | $ | – | $ | 3600.00 | $ | 7200.00 | $ | 10,800.00 | $ | 14,400.00 | $ | 18,000.00 |
$ | 2,614,000.00 | $ | 13,519,085.71 | $ | 24,424,171.43 | $ | 35,329,257.14 | $ | 46,234,342.86 | $ | 57,139,428.57 |
Applications | Private Cloud | Public Cloud | No Change |
---|---|---|---|
Migrate as-is | 97 | 49 | |
Transform or Re-platform | 71 | 17 | |
Re-install | 36 | 27 | |
204 | 93 | 53 |
Application portfolio assessment
-
SAP – The TPR is Private Cloud and is assessed as suitable. As the SAP version is current and deployed to VMWare, the decision is to Migrate As-Is to Private Cloud.
-
Contact Centre – The TPR is Private Cloud and is assessed as suitable. The version of the software is 2 versions behind the current version; hence the decision is to transform the software through a two-stage upgrade process.
-
Microsoft Outlook – The TPR is Public Cloud and is assessed as suitable.
-
ESRI – The TPR is Private Cloud and is assessed as suitable. The software will be re-installed in private cloud with cost effective run time costs due to the static nature of the application
-
Inventory - The TPR is Public Cloud and is assessed as required to be in Private Cloud due to data dependencies and integration with SAP. The target platform recommendation is changed to Private Cloud and the FVA re-run.
-
Wave 1: In the first wave, 151 applications comprising Web Servers and applications running on VMWare are selected for migration with the duration estimated to be 1.5 months.
-
Wave 2: In the second wave, 67 independent applications that require transformations are selected. The application transformation is estimated to take 3 months and the application migration is estimated to take 2 months.
-
Wave 3: 54 applications with dependencies are grouped together for migration in Wave 3 with a subset requiring transformation. The application transformation is expected to take 2 months and the application migration is estimated to take 3.5 months.
-
Wave 4: 25 applications with dependencies are grouped together for migration in Wave 4 with a subset requiring the most complex transformation. Application transformation is expected to take 4 months while migration is estimated to take 4–5 months.
Recommendation
-
Q1: Which applications are suitable for Cloud Computing?
-
297 out of 350 applications in the portfolio are suitable for Cloud Computing
-
Q2: What are the cost savings of moving those applications to the Cloud?
-
The cost savings of moving those applications to Cloud Computing is $350,821/month. Additional savings of a reduction in two people in the platform support team due to a smaller footprint in the bank’s data centre has resulted in a saving of $350,000/month or $29,166/month.
-
Q3: What is the split of applications being deployed between public and private cloud?
-
93 applications are targeted for public cloud (31%) while 204 applications are targeted for the private cloud. 53 applications are left as-is.