IT organizations are consistently under pressure to do more with constant or declining resources. IBM PureApplication System allows IT organizations to work smarter and become more efficient with time, skills, and resources.
PureApplication System can vastly simplify or eliminate many of the routine tasks that consume the time of development, operations, and system administrator staff. Freedom from these repetitive tasks creates a unique opportunity for businesses to fully leverage the deep technical skills of its organization by assigning them to higher value tasks.
PureApplication System utilizes roles an IT organization is already familiar with such as security, hardware, and cloud administration. It also introduces new concepts and roles such as patterns and workload administration, which will require a combination of skill sets that are likely already found within the IT organization. Management can gain maximum benefit as they promote these new skills.
This article is written for IT managers and CIOs who are looking to understand how PureApplication System can affect the existing roles within their organization. We discuss how an IT organization using PureApplication System can optimize utilization of existing skills and leverage its current resources to fully realize the benefits of an expert integrated system. IT organizations will discover that PureApplication System can promote more collaboration between application development and IT operations. By making changes to roles and remixing skills, PureApplication System allows these two teams to react faster to today's rapidly changing business environment. This article provides examples, using data from commissioned studies, results from IBM Services engagements, and client discussions, to show how some existing IT roles can be affected by PureApplication System. Suggestions are made on how to map to roles that will emerge in the new environment.
For more information on integration by design and patterns of expertise, refer to the IBM PureApplication System web site.
This section introduces the benefits that can be derived by realigning the organization after implementing PureApplication System.
Traditional IT environments are complex and encourage specialization of IT roles. This specialization often resulted in a tendency to compartmentalize the IT organization into teams responsible for very specific tasks. While this was done to optimize the performance of applications or specific infrastructure components, the result was often an organization that inhibited collaboration between IT operations and software development. Rapidly changing business environments are driving for greater collaboration between operations and development, creating the requirement for an infrastructure that can facilitate this operational model.
With patterns of expertise providing optimized middleware stacks and pre-integrated hardware simplifying infrastructure delivery, a small operations staff will be capable of supporting the entire breadth of hardware and software required to implement a single application. Such an operational model provides the development team with a point of contact who can manage all components of the infrastructure that its application depends on. This leads to smaller teams who can work more efficiently on tasks such as problem resolution, which would have required larger teams to obtain and interpret the data, identify and fix the problem. As these teams work together, they should find planning and implementing projects can occur with fewer delays.
Highly skilled people in IT roles commonly perform unskilled tasks for a significant portion of the day. This situation is nearly unavoidable with the typical compartmentalized structure where a job role focuses on a specific technology. For example, an IBM DB/2® specialist may have rare skills and knowledge on how to tune a database to work correctly with transactional workloads. However, the specialist will also have to spend a significant amount of time installing the software product and setting up user IDs, connecting it to an enterprise directory, and configuring instances.
PureApplication System addresses this problem for IT organizations because it helps to automate most of these repetitive tasks. This helps free up the specialists to spend more time using their deep skills to benefit the business. The IBM WebSphere® Application Server specialist, for example, spends more time helping developers troubleshoot a performance problem with an application under development, or helps an architect design a new application pattern to be included in a catalog for the entire organization to use.
Pattern-based deployment introduces the need for individuals who are specialists in the implementation of patterns. These specialists have deep knowledge in pattern design, configuration, and integration, and own a pattern from the initial requirements phase through architecture, development, test, and into production. They can consult with application owners on requirements and best practices to assure new applications are compatible with and make best use of the pattern.
PureApplication System shifts the model for requesting resources for new applications from procuring new to allocating existing resources. Resource specialists have an overall view of resource utilization and can communicate with new application owners to determine the best options for CPU, memory, and storage and the optimal scaling policies to assure the application meets performance goals.
PureApplication System combines the major parts of a large-scale, virtualized, distributed system into one package. Tasks that used to take teams of specialists days or weeks to complete can be accomplished by a wider range of people in much less time. This is accomplished by capturing the deep skills of IT specialists inside the system and into the deployable patterns. The knowledge of how to tune a complex piece of middleware or how to scale an application for a specific kind of workload can be encapsulated so that anyone with lesser skill in the organization can take advantage of it.
These patterns allow the user to take on roles that cover a broader range of tasks. For example, WebSphere Application Server administrators no longer have to spend time tuning their application servers because that task is being managed by policies defined in simple business terms within a pattern. The administrators can instead branch out and apply their skills to assist application developers and architects who are implementing their patterns. This model provides a vehicle for operations and support staff who aspire to expand their expertise and have greater overall influence on an end-to-end application.
The technology delivered with PureApplication System makes existing IT personnel more efficient, providing the IT organization with a unique opportunity to focus valuable skills on addressing new business opportunities. This section explores several areas where integration and automation can help change the way that individual specialists interact with the system.
Before PureApplication System, a dedicated group was required to manage the storage area network (SAN) infrastructure. Apart from setting up the physical components of the SAN, someone had to be responsible for the SAN and LAN planning to connect the SAN infrastructure to your hosts and to perform such tasks as host and volume configuration, host mapping, and storage pool management. Likewise, after the infrastructure was set up, someone had to constantly care for and manage it to keep up with new driver levels, SAN software levels, and so on. PureApplication System does this for you. Often, the only storage tasks that might need to be performed during normal operations are decisions about creating the appropriate volume size when a new virtual machine is provisioned and expanding volume sizes when needed.
Connecting traditional distributed systems to a network involves a mass of cabling and considerable planning. Wiring the system components into network switches and storage arrays consumes considerable time and is error prone. After the physical setup is complete, the task of assigning IP addresses and hostnames and creating VLANs begins. The end result is a physical and logical configuration that must be managed and maintained, and this system is typically fragile with respect to change. PureApplication System arrives prewired for network and storage and for help with reducing the setup to a set of configuration steps that can be completed quickly. Maintaining the system is similarly done through logical configuration. There is generally no need to pull and move cables to reconfigure the network. As a result of this approach, fewer networking specialists are required to build and maintain a PureApplication System when compared to traditional distributed hardware.
PureApplication System changes the strategy and approach for implementing a virtualization infrastructure. Building a new virtualized environment traditionally requires someone who is responsible for installing hypervisors, installing the virtualization management infrastructure, connecting the individual hypervisors into the infrastructure, and then maintaining the assembled infrastructure. PureApplication can help provide this extensive functionality, and can play a role in making the virtualization specialist role redundant and individuals assigned to this role available for reassignment. Specialists who are skilled in system virtualization can take on the role of PureApplication System administrator, with the tasks of defining and managing virtualized resources.
A large amount of project start up time is spent in resource acquisition, installation, and configuration. New projects typically require the procurement or allocation of hardware, data center networking, and other resources. After hardware is in place, operating systems need be configured and middleware installed. Everything within this startup phase must be completed before development on the project can begin. We found this can take on the order of ninety days. When Pure Application System is purchased, you may find that the need to acquire hardware on a per-project basis turns into a request for resources on the existing system. In such circumstances, hardware specialists who previously built the traditional infrastructure can take on administrator roles responsible for creating and maintaining the individual environments within Pure Application System.
These general areas demonstrate where PureApplication System can be used as a tool to address the workload of individuals responsible for carrying out specific tasks. Contributing to reduced resource demand, an organization may find that these individuals can take on higher business value tasks.
This section provides examples of how PureApplication System can help reduce the workload on traditional IT roles and tasks. We then look at how existing resources are mapped to new roles that will emerge as part of implementing an integrated system.
For a sampling of existing IT roles, Table 1 summarizes typical responsibilities, and the estimated potential reductions in the effort required for those responsibilities after PureApplication System is implemented.
Table 1. Effect on existing IT roles
|Role||Responsibilities||Estimated reduction in workload|
|Middleware administrator||Maintenance, performance tuning, monitoring, problem determination||60%|
|Middleware architect||Application integration, application performance tuning||30%|
|Hardware administrator||Hardware planning, hardware utilization management, hardware assembly||90%|
|Database administrator||Database planning, database installation and configuration, database management||60%|
|Network administrator||Network configuration||50%|
|Application owner||Architecture validation, product license management||60%|
|Procurement||Hardware and software procurement||90%|
Note: The estimated reduction in workload is based on data obtained from client workshops.
For each typical IT task, Table 2 depicts a typical duration before PureApplication System implementation and time savings expected after PureApplication System implementation. The list is not comprehensive and serves as an example.
Table 2. Expected impact of PureApplication System on work effort
|Steady state tasks|
|Product maintenance||Implement product upgrades, patches, and fixes||52||75%|
|Application performance tuning||Monitor and manage application performance parameter/configuration settings||18||25%|
|Product performance tuning||Monitor and manage middleware performance parameter/configuration settings||52||75%|
|Problem determination||Product and/or infrastructure problem determination and troubleshooting||78||50%|
|Application management||Application management and configuration support activities||15||50%|
|Product troubleshooting||Identify and remediate problems with runtime environment.||52||50%|
|Process management||Start/stop procedures for middleware components and operating systems||20||75%|
|Procurement||Procure hardware for project||43||100%|
|Procurement||Procure software for project||52||75%|
|Install hardware||Install hardware, cabling and network||46||90%|
|Configure environment||Tune environment in preparation for application deployment||52||75%|
|Application integration||Integration, configuration and testing of middleware stack||55||75%|
|Product monitoring||Design and implement specific product monitors based upon application requirements||59||50%|
|Application resiliency||Implementation of clustering and HA environments||58||50%|
Note: The task duration estimates are based on commissioned studies and customer workshops. The potential savings are based on data obtained from client workshops.
The reduction in effort described in Table 2 frees resources to take on new tasks that arise with the adoption of PureApplication System.
For each new role in a PureApplication System implementation, Table 3 identifies the role, summarizes the role responsibilities, and identifies the traditional roles that are contained within the role.
Table 3. Mapping existing IT roles to application aligned tasks
|Role||New responsibilities||Existing roles|
|Application Software Architect||Responsible for middleware selection and for overall systems design of applications||System administrator, database administrator, middleware architect|
|Pattern Developer||Codify existing environments into patterns Deep knowledge of implementing patterns. Consult application owners on best practices for utilizing existing patterns. Develop scripts to extend functionality.||Middleware architect|
|Pure Application System Administrator||Manage system configuration and resource utilization (compute resources, cloud groups, storage allocation)||System administrator, hardware administrator, network administrator, middleware administrator|
|Asset Librarian||Manage patterns, scripts, and images in the asset catalog||System administrator|
|Physical Cloud Architect||Configure and manage the physical cloud assets in the PureApplication System. Create isolated clouds within the system.||System administrator|
|Application Cloud Administrator||Creates and manages environment profiles. Maintains users rights to environments.||System administrator|
After implementing a PureApplication System solution, the requirement for new skills will emerge. Architects, administrators, and developers are candidates to acquire these new skills and take full advantage of the new technology. This section describes some of the new skills that will be required to implement and manage a PureApplication System.
This section introduces the new skills needed when implementing PureApplication System.
To maximize the value that can be achieved by implementing patterns, organizations need to analyze existing workloads, identify similarities, and group applications into categories. These categories of like applications can be the basis for developing useful patterns. For example, an organization may find that they have a large number of applications that consist of a web server that serves the user interface, an application server for application logic, a database that stores persistent application data, and a user registry that is used for application authentication and authorization. The integration of these four components is similar across all applications, making it a candidate to be implemented as a pattern. Existing system administrators and architects who are skilled in the design, integration, and configuration of distributed applications are well positioned to pick up the skills required to begin pattern identification and design.
Application patterns typically require some degree of customization to meet the needs of the application that is being deployed. Application specific details such as database name, security roles, and runtime parameters vary from application to application, even though they are implemented using the same pattern. To address this requirement, two options are available for making the necessary adjustments:
- Post-deployment scripts that modify parameters each time a pattern is deployed.
- Extend and capture, which modifies the existing pattern and creates a new copy with the changes.
Both approaches require the skill to understand existing patterns and application requirements then identify the best solution. This skill set does not exist outside of a platform as a service environment and is a good match for current system administrators who understand the requirements that applications have on their environments.
Individuals who are familiar with the design point of patterns can consult with application owners on which patterns best fit their application needs. They can also advise on how to properly design new applications to make the best use of existing patterns. These individuals have end-to-end knowledge of all components within the patterns, as well as all interactions, configurations, and dependencies.
PureApplication System ships with a catalog of virtual application components representing the most common enterprise infrastructure components. In addition, IBM provides a Plug-in Development Kit that allows customers to build required components that expand functionality beyond what is delivered with the product.
This section introduces the skills required to manage PureApplication system.
Physical Cloud Administrators configure and manage the physical cloud assets in PureApplication System. They create new isolated clouds with the desired quality of service as needed for different internal or external groups who use the system, manage shared services in the cloud as required by applications, and manage and monitor cloud resources.
The emerging role of Application Cloud Administrator creates and manages cloud environments and maintains the users that have rights to deploy and manage applications running within the environment.
This role manages the catalog of assets, which includes images, script libraries, and patterns. They are responsible for asset lifecycle across the organization's set of PureApplication Systems.
PureApplication System allows for environments to be partitioned quickly and easily, thereby generating smaller and shorter-lived middleware environments. Environments that are difficult to create tend to grow to accommodate many projects and exist for an extended period of time (one to four years). Patterns provide the ability to provision an environment within minutes of when it is requested and to deprovision the environment when it is no longer needed, thereby freeing resources for use by other environments. Users are more likely to remove unused environments when they know they can be quickly re-created. This approach also leads toward more small development and test environments. Each application project can have its own contained space without impacting applications that are under development at the same time. Administrators spend less time building and maintaining large environments and more time developing scripts that facilitate the rapid and repeatable configuration of provisioned environments.
Typical application serving environments consist of numerous components such as hardware, operating system, network, application server, and database. Each of these components can be maintained and administered by different groups or individuals, such that a single administrator rarely has the insight or ability to affect all aspects of the environment. By managing most of the underlying infrastructure and exposing only administrative features necessary for day-to-day operations through a unified console, PureApplication System makes it possible for a single or small group of individuals to administer all aspects of a single application.
This section reviews organization and process changes that commonly emerge after PureApplication is implemented.
The complexity inherent in building large IT environments encourages specialization. To leverage the pre-built and pre-integrated solution patterns from PureApplication System, organizations have the opportunity to integrate and work more closely throughout the IT lifecycle.
The reporting structure in today's typical IT organization is complex (see Figure 1), with multiple layers compartmentalized to work with the existing technologies. Often, the faster the rate of change in the external environment (that is, IT technologies and capabilities), the more the organization structure changes to adapt and leverage the new features.
Figure 1. Typical reporting structure
Figure 2 shows how the reporting structure changes when PureApplication System is implemented.
Figure 2. Potential reporting structure with PureApplication System
With PureApplication System, the IT reporting structure may become more optimized, simplified, and streamlined with less management interference.
Organizational change should improve the performance of the organizations and their people. Therefore, understanding what is meant by performance and the methods to manage performance is important. One method is to readjust or re-engineer IT lifecycle activities and processes to leverage PureApplication Systems. Some tasks or activities in the existing process might be eliminated, reduced, or modified. Process changes typically result in an organization that is more productive, dynamic, and faster to respond to changes in market demands, business requirements, and technologies. Figures 3 and 4 show how the complex and inflexible IT lifecyle process typical today can be transformed to a productive, responsive, and dynamic process with fewer steps and faster completion times.
Figure 3. Today's complex and inflexible IT lifecycle process
Figure 4. Transformed to a productive, responsive, and dynamic process with fewer steps and faster completion times
Another consequence of the less compartmentalized organization is that governance processes change. Because the role of individual groups such as networking and systems management are reduced, there is less opportunity and need for them to execute their own individual governance processes. Instead, what may be more appropriate is a centrally defined process executed by groups of generalists. Again the issue is that with expert integrated systems, everything happens faster and is driven through modeled approaches such as IBM-provided hypervisor images and virtual applications.
IBM PureApplication System is a new offering that leverages integration by design and built-in expertise to help reduce the cost and complexity associated with implementing and managing an application serving environment. It can be used as a tool to reduce time to develop and test applications while helping improve consistency through automated deployment of applications using pattern-based models. Policy based management of the application infrastructure may reduce administration costs, allows auto scaling, and can be tied to business requirements.
Consistent and simplified administration and deployment have the potential to reduce labor costs by estimates of more than 50%. Elastic resource management adapts to changing business needs and can help lower total cost of ownership.
The IT organization (skills, roles, and team structure) has the opportunity to change and take advantage of the potential for labor cost reduction with the aid of PureApplication System. IT leaders can help drive the strategic transformation of their resources with approaches by considering the following steps:
- Streamline the existing IT processes to leverage the new capabilities in the invested technology.
- Redefine roles and retool existing skills to take advantages of new integrated and automated processes.
- Transform the IT organizational structure to be more integrated, dynamic, and responsive and to employ cross teams and cross disciplines.
- Focus on business value and the team potential with the new invested technologies to establish guidelines and policy.
- Track actual progress against expectations and calibrate change where necessary.
The authors would like to thank the following individuals for their contributions in providing ideas and content and for their reviews of this article: Andrew Hately, Greg Boss, Manuel Silveyra, Rejeev Gandhi, Susan Holic, Hans Skalle, and Beatriz Friedman.
Preparing for IBM PureApplication System, Part 1: Onboarding
Preparing for IBM PureApplication System, Part 2: Is your application
ready to become virtual?
IBM WebSphere Developer Technical Journal: The Ideal WebSphere
Collaborative development and operations (DevOps) page
Create and customize virtual application patterns
Redbook: Implementing the IBM Storwize V7000 V6.3
IBM PureApplication System web site
PureSystems resource page on developerWorks
Vincent Belfoure is a Cloud Architect for IBM Cloud Labs, a division of IBM Software Group. He has been with IBM for 17 years and currently works with customers in designing cloud computing solutions. Prior to joining the Cloud Labs, Vincent held positions as the BPM Technical Sales Lead for Central Eastern Europe, Middle East and Africa, and Worldwide Technical Sales Lead for WebSphere Process Server. He has also been a member of the New York City WebSphere Technical Sales team and IBM Software Services for WebSphere.
Kyle Brown is a Distinguished Engineer with IBM Software Services for WebSphere and specializes in SOA and emerging technologies. Kyle provides consulting services, education, and mentoring on SOA, object-oriented topics, and J2EE technologies to Fortune 500 clients. He is a co-author of Java Programming with IBM WebSphere and Persistence in the Enterprise. He is also a frequent conference speaker on the topics of SOA, Enterprise Java, OO design, and design patterns.
Jeff Coveyduc is a Program Director in IBM Cloud Labs. He leads Cloud Labs teams at IBM Silicon Valley Laboratory in San Jose, California and across the globe, working with enterprise customers around their cloud computing needs and adoption. Jeff has been responsible for successful enterprise engagements and references spanning sectors from financial services, to telecom, healthcare, and public sectors. Since joining IBM in 1999, Jeff has held multiple positions within IBM's Software Group including WebSphere, Tivoli, and Strategy and Technology. Jeff has been recognized for his work in growth markets, establishing the IBM Africa Innovation Center and Software Solutions Laboratory in Johannesburg, South Africa, as well as opening satellites in Kenya and Nigeria while on assignment to sub-Saharan Africa.
Son Huynh is a Cloud Solution Executive for IBM Cloud Labs, IBM Software Group, and has been with IBM for 18 years. He is also a member of IBM Cloud Academy (ICA) and its Technical Architecture Board. Son leads a team that develops and delivers a Business Value Assessment (BVA) process and Organizational Change Management for new IBM Cloud and IT solutions. Before joining Cloud Labs, Son held positions as the Solutions Executive for Chemical and Petroleum Industry in IBM Industry Solution Group. Before IBM, Son was the Product Manager of Chemical and Pharmaceutical Industry Group for SAP, Germany. He also served as Integration Technology Manager for the Monsanto Chemical Company in St. Louis, Missouri prior to working for SAP.