Agile Modernization Framework for Decision Making in Legacy Environments

Development of enhancements in a legacy environment poses a major application development challenge to many finance and insurance companies. Investment in modernizing architectures does not always immediately realize operational and customer-perceivable value mandated of agile environments.

And top-down service portfolio management does not consider architecture impact on decision making to make the priorities and investment realistic. To address these challenges we provide the Agile Modernization Framework (AMF) for systematic decision-making that includes bottom-up architectural and infrastructure information along with top-down and strategic customer and operational needs to make business decisions.

We show how this unique hybrid prescriptive framework 1) guides the decision making of legacy-related application features, 2) aligns and constrains implementation in the context of an agile business environment and strategy, and 3) applies to realistic examples where services are often based on shared components and numerous services are involved in complex interrelationships. Categories and Subject Descriptors D. 3. 3 [Programming Languages]: Language Contructs and Features – abstract data types, polymorphism, control structures. This is just an example, please use the correct category and subject descriptors for your submission.

The ACM Computing Classification Scheme: http://www. acm. org/class/1998/ General Terms Management, Measurement, Documentation, Performance, Human Factors, Standardization, and Verification. Keywords Legacy portfolio decisions, decision-making metrics, modernization, migration, IT service portfolio management. INTRODUCTION We introduce the Agile Modernization Framework which supports a “stakeholder driven & legacy impact based” decision making process for the service portfolio of the enterprise. The overall benefit is to align Agile development practices with value and agility benefits at the Strategy, Business & Operational levels.

Problem: In large financial and banking institutions, while most have adopted ITIL and an IT services portfolio, they still have a large application backlog and significant complexity due to business-critical legacy systems. With a large legacy application portfolio, providing services to stay competitive within profitable margins has become increasingly difficult because of associated maintenance costs. However legacy systems continue to provide core services to organizations, handling the organization’s daily operations. The organizations are also heavily dependent on these systems as they contain the knowledge accumulated over years.

The problem we address is providing decision makers with a clear and comprehensive overview of tradeoffs with their existing legacy application portfolio, that will form the basis for making strategic high level management decisions, that realistically aligned to the challenges faced in a legacy environment. For this type of decision-making, IT executives in an agile environment need to know how much each of those high-level services costs to support, manage and operate with respect to the value each one delivers to the company and operational users.

This is particularly challenging because it needs: 1) Analysis of tangible and intangible value indicators for the Business, Infrastructure (information technologies, facilities, human etc. ), Operational, and Strategic (BIOS) dimensions of stakeholders in the organization. 2) Provision of an architectural overview of an extremely complex system, which involves handling substantial number of performance uncertainties and assumptions related to new development. 3) Defining architectural impact on the operations. 4) Risk estimation of the investments in legacy applications.

Integrated top-down/bottom-up solution: We provide a framework that serves as the guiding method that all the dimensions of organizations and resources can rally behind for integrated top-down and bottom-up planning. It helps negotiate conflicting priorities. It’s not a project plan per se, but it summarizes at a high level the prioritized services, their dependencies and pre-requisites and the time frames that they may get worked on. Our framework effectively encapsulates the interests of each dimension of the BIOS stakeholders using feature metrics/ indicators and then determines the decision zone of implementation approach to be followed.

These zones provide options ranging from completely changing the application which could involve discarding the existing application and writing a new application from scratch or investing in third party solutions; to sticking with the existing status quo (maintain); or to improving the application with an Service Oriented Architecture-style(modernize) and could involve application to run on a new platform(migrate). This zone is determined first for that service (henceforth we use this term generically to refer to service, feature or component) in isolation; second we consider the service with respect to the other services.

The Agile Modernization Framework (AMF) provides the BIOS decision makers and service owners with a service graph-based method as follows: 1) Strategic and Business owners (S and B dimensions) determine the existing and proposed highest-level services and their customer facing and business priorities and pass them to the Infrastructure developers and Operational users of those services (I and O dimensions); 2) for each sub-service, the indicators of interest to the I and O dimensions and also reflective of the legacy challenges are determined; 3) the indicators of each sub-service and/or shared service are next propagated up to the service above and synthesized. Thus the overall influence of one service on others is considered.

We will show through a real-world application (abstracted here for space considerations) how this bottom-up synthesis sometimes results in a more dramatic impact to the portfolio and might cause rethinking of investments and even prioritization (e. g. a decision to take a modernization approach and move from file systems to a data base because many features could share the technology and benefit). AMF-based IT service portfolio management thus involves not only the top down initiatives in the enterprise but also addresses the various interdependencies and complexity between various services at lower levels and thus injects the reality to properly set expectations upwards. Figure [ 1 ]: Service and decision zones.

Target Audience: Unlike portfolio planning frameworks intended only for higher-dimension decision makers (typically CxOs, business/strategy analysts, enterprise architects), the AMF is intended for collaboration between high-level decision-makers and service/application professionals (software engineers, resource planners, etc. ). Paper outline: Fill in. RELATED BEST PRACTICES AND RESEARCH Best practice related to legacy environments is explicitly covered, as factors leading to challenges in the first sub section. This is used to derive some of key indicators unique to this environment. In the next sub-section we cover related work in Application Portfolio Management. Next we cover the body of work done in ITIL and some of the standard terminology used in AMF. 1. 1 Stakeholders Interests Related to Legacy

There are many tangible (measureable metrics), non-tangible (soft) and risk factors associated with decision-making criteria of managing and legacy applications. Based on our research in Business – IT alignment [2], we consider the four stakeholder dimensions identified below. We also identify the BIOS causal chain at a high level to be “Business invests in Infrastructure enabling Operational changes that contributes to Strategy”. Note that these dimensions are not only related to the Balanced Score Card [] but also include an additional dimension and indicators that identify the complexity of the IT architecture. This will allow planning participants to view the tradeoffs from multiple different perspectives.

Each of these dimensions has identified metrics and indicators of interest to its stakeholders: Business dimension: The Business stakeholders typically deal with the costs and time associated with the development /maintenance for legacy applications. Total cost of ownership (TCO)[30], Return on Investment (ROI), Internal rate of return or payback period are the major value indicators for these stakeholders[10]. Infrastructure dimension: These stakeholders consist of the IT resources of the organization – personnel, hardware, and software acquisitions/ upgrades/ maintenance. Interdependencies between different components/resources etc. , people skill set, complexity of the workflow and code coupling within the service are some value indicators here.

Operations dimension: This deals with the internal customer user aspects of the particular service/feature. Usability to deliver on end customer needs efficiently is an indicator. Strategy dimension: The strategy stakeholders create the business case for a particular service/feature. These indicators drive the top-down flow in decision-making. Indicators include: increase in customer satisfaction and # new customers. Customer perception and service scalability are also other major indicators. 1. 2 Overview of Application Portfolio Management (APM) & Legacy Challenges Many large-scale enterprises have amassed a complex infrastructure of applications, which includes legacy systems.

Application Portfolio Management (APM) is designed to help an organization transform its application portfolio to more closely align to the business and responds quickly to business change. APM consists of the IT portfolio of prioritized projects and the process of continually assessing and evaluating priorities based on balanced business and technical ROI criteria. The management decision-making includes alignment of the portfolio with the strategic needs of the business and supporting its transformation. [40] The APM process documents the business processes and its application dependencies. This is through both bottom-up discovery and manual top down surveys, documentation reviews and interviews, which are then consolidated.

This is then recorded in the single portfolio repository, enabling comprehensive visibility and control of the application portfolio. APM enables the organization to assess and prioritize the portfolio for rationalization into zones. For example any modernization would be based on alignment of business goals with IT technology decisions. The output of the assessment is a strategic road map for portfolio project implementation and transformation. Managing the transformation not only drives the implementation as prioritized by the road map but also specifies the ongoing transformation steps, the governance, and the metrics collection so that business value indicators can be derived to ensure identified objectives.

It is also vital that the portfolio is reassessed periodically to ensure it continues to support changing business priorities. The APM process consists of three activities identified next [38]. Metrics Data Collection: In the Data Collection phase, the current elements belonging to the application portfolio are captured. Next ‘applications rationalization’ is used to assess the TCO and ROI of applications. Applications Rationalization [39] is a process that inventories the applications by assessing cost, quality, risk, TCO and business value to define a future state consistent with the business and IT strategies and develop a transformation roadmap of actionable projects to realize that future.

ROI is evaluated in terms of strategic alignment, technical quality and business quality. Analysis of applications also includes risks that focus on the probability of application failure and the effects this failure would have on systems and business. Note that AMF adds to typical portfolio management practice by emphasizing that to gain comprehensive insights, it is not sufficient to analyze the portfolio by Business indicators alone. Rather the aggregated BIOS impact of legacy considered in this paper is needed to conduct a more thorough assessment and risk analysis. Decision-Making: The second step in the APM process is planning and shaping the to-be portfolio.

It is based on detailed application analysis and, in general, involves determining how to allocate investments in existing applications, the purposes of such investments, and which new applications or infrastructure warrant investments. Optimization: The Optimization phase puts the results of the Decision-Making phase into action. Challenges of using APM in Legacy Environment: One main challenge of using APM in legacy environment is that the scope of a portfolio is not standardized and is subjective. To consider interferences across applications, we must look at all the applications. Here we introduce the notion of ecosystem later, for this purpose.

Also legacy duplicate applications are a result of mergers or acquisitions, which increase the cost and complexity of implementing new requirements. Past pressures to deliver solutions rapidly have created a legacy of tactical solutions rather than a structured application portfolio. Additional interfaces to support these solutions add yet more complexity to a unstructured applications architecture. This can be due to highly decentralized IT management, or the existence of various legacy applications that have not been migrated. A condition called ‘applications sprawl’ which means a portfolio of that was built up over time, across business units and geographies is now out of control due to growth and acquisition [41].

In complex legacy application environments, it is usually impossible to view each application in isolation. There are tight interdependencies and couplings amongst applications. Consequently, interconnected applications and their analysis results need to be viewed together as a whole that is as a single portfolio. For instance, a “rehosting” analysis result for one application may not make sense if the analysis result of another application says “retain” especially when both applications are tightly coupled and provide another shared service. In AMF we provide a solution for challenges such as these by defining a comprehensive services portfolio in such a way that it encompasses all the interdependencies.

For ITSPM [14], the amount of legacy information to be examined and analyzed is growing at an exponential rate. Corporations have installed software to assist in organizing information on to solve this problem. However, for each silo of information, the software tool chosen is usually platform-specific and none of these tools was designed to communicate with other areas in the organization where related information exists. The time and effort spent in information gathering is directly related to the accessibility of information. A typical analysis will need to gather information either in raw format or in the proprietary format which is kept in people’s heads, in hardcopy only, on the mainframe or on the PC, in various formats [42]

In summary, the proposed AMF addresses disadvantages of the strictly interpreted APM that fails to evaluate the services of complex systems of applications based on indicators of interest to Business, Infrastructure, Operations, and Strategy stakeholders. 1. 3 IT Service Portfolio Management The first mention of the portfolio concept as related to IT was from Richard Nolan in 1973: “investments in developing computer applications can be thought of as a portfolio of computer applications. ” [Nolan, Richard (1973)]. In the widely employed ITIL IS0-20000 best practice, IT service portfolio is defined as “a complete list of the services managed by a service provider and is divided into three phases: Service Pipeline, Service Catalogue, and Retired Services.