Cloud

Optimization and Decision Support for an Enterprise
An Application of Cloud Computing

Gary S. Schebella/ NBS/ (703) 999-1849

1.0        Overview

Enterprises have a need to consolidate their disparate information technology (IT) applications and services in order to augment ease of use and
to reduce costs. One approach is to change the physical locations of data and processors so that access times are minimized and system security
is enhanced. An alternative is to create a private cloud environment for central facilities and a public cloud environment for distributed assets. A
cloud methodology is initiated with a mapping of a cloud to an enterprise, and subsequently to consolidate all resources through the process of
virtualization. Applications of private and public clouds are now state-of-the-art. Effective security practices are also a part of the cloud
methodology.
But, within a cloud, every data source might be in a constant state of flux. Even so, additional enhancements to data and enterprise management
are achievable in the form of a decision support system (DSS). The DSS provides management courses of action without disrupting current
procedures.
NBS Enterprises (NBS) is foremost in the development of management decision aids. We are able to associate dashboard data from disparate
sources such as databases for program costs, resources, contracts and schedules. For example, if a new project is proposed as an addition to a
program and no additional funding is available, a manager is interested in the project impact upon schedule, priorities and risk. Answers to this
query, by a staff operating without automated decision support, require several days. The NBS decision support system (DSS) associates the
dashboards and provides answers within hours or minutes. Further, NBS has developed a set of optimization algorithms that assist with planning,
forecasting and prioritization. Because our algorithms are fully developed, their applications produce great savings in costs and time of systems
development. With a DSS, a cloud environment becomes more than a receptacle for data. The cloud is transitioned to a mechanism for constant
research, experimentation and analytical forecasting.
2.0 Problem Statement
An organization is responsible for enterprise management and decision making. Multiple variables are considered with disparate databases as
there sources. Manual processing is time consuming and costly in terms of labor.
Thus, the effectiveness of managers, when associating multiple variables and developing courses of action, is improved dramatically with state-of-
the-art automation techniques that produce optimal courses of action.

3.0 Solution

A decision support system is ensconced in the framework of a cloud environment.
Following and during cloud development, an incremental encapsulation of a DSS is initiated whereby an architectural and enterprise analysis is
followed by design and development.
After the completion of analysis, a development schedule is provided for management review. The second phase of development establishes a
capability to associates disparate databases and an analytical forecasting capability. The tools identify future requirements and are responsive to
asynchronous change. The third phase of a project applies NBS optimization algorithms to a total enterprise. The algorithms provide guidance on
metrics and measurements and insert their results into the DSS.
The optimization computations are always entirely visible to a management staff and are responsive to queries and “what if” analysis.

3.1 Operational Procedures

After data are extracted from disparate sources, the existing decision support system (DSS) provides computations which represent enterprise
performance. A metric might be customer satisfaction with respect to staff workload and the costs of operations. A more comprehensive set of
metrics provides not only a high-level statistic of an enterprise, but also statistics that relate to the performance of the DSS itself. Team NBS has
developed an analytical tool suite that represents, evaluates and optimizes distributed enterprises and systems. Applications of the tool suite assist
with the prioritization of resource allocations and the management of an enterprise, as well as measurements of performance for services. The tool
suite has been well tested and has been applied to several decision support initiatives such as program management, political analysis and the
deterrence of terrorism. For information technology (IT) systems, we have addressed the Reserve Component Automation System (RCAS), the
Saudi Air Defense System, the Comanche Helicopter, NASA Financial Management and Space Station Software development. In every instance we
have produced exceptional results in the conversion of “as-is” to “to-be” system architectures and designs. We always reduce development and
operational cost. For protective or battlefield missions, lives also might be saved. Thus our tool suite is labeled “Time, Lives, Cost” or TLC. The
TLC mission-planning tool suite is instrumental in the understanding of enterprises, distributed systems and diverse technological components, as
well as their performance relative to a system administrator’s objectives. TLC capabilities include
a)        Reasoning with uncertain information
b)        Representation of data at different levels of granularity
c)        Rapid capture of changes
d)        Analytical forecasting and automated courses of action to assist with planning decisions
e)        Technology forecasting

To fully understand the impact of operational functions and the dynamics of an implemented or proposed system, modeling and simulation are
necessary to compute performance statistics for all of their artifacts. Further, the impacts of current and forecasted technologies need to be
assessed. Numerous tools and simulation languages are available for applications. However, a COTS tool which encapsulates the capabilities to
generate both performance calculations and the optimal disposition of system components is not available on the open market. TLC not only
derives performance statistics and optimizes the disposition of resources, but it also provides analytical forecasting and impact analysis.
Applications of the TLC produce the answers to system administrator questions, namely, 1) “How well does the enterprise, its systems and its
components function (performance analysis)?”; 2) “How can we improve operations (optimization)?”; 3) “What is the value of a new technology (
component performance in the context of the enterprise)?,4) What is the risk of a program change?, 5) What are overall risk factors and how can
they be mitigated?, and 6), “How do we plan for the future (analytical forecasting and the transition from as-is to to-be)?”.
In essence, qualitative representations are transferred to quantitative models, expediting analysis and producing a rationale for change.

4.0 DSS Capabilities

A baseline system establishes a single computer server with a cloud environment that captures all information relevant to enterprise management:

a) Provide a capability to view a system architecture.

c) Provide the capability to make changes in a concept of operations.

d) Provide analytical forecasting of resource allocations and planning support.

e) Provide plans responsive to asynchronous change.

f) Provides answers to queries, “what-if” analysis and the association of planning factors such as costs, schedules, resources, and technologies.

Appendix A: Cloud Computing

A 1.0 Cloud Computing

Cloud computing comprises a set of pooled resources delivered over a network. It provides a host environment that is not limited to a specific set of
resources and it can be expanded dynamically to multiple users.
A private, or internal cloud, applies the concept of on-demand resources that are owned wholly by an enterprise while providing security and
regulatory compliance.
External, or public clouds, involve IT resources and services with on-demand provisioning. They are accessed with web browsers and offer large
capacities dynamically.

A 2.0 Architectural Analysis and the Mapping of Clouds
Various stages of architectural analysis are required for the initiation of cloud computing. In essence, an understanding must be gained of an “as-
is” system with respect to users, communications, processors, data, etc. If the system is to be changed, qualitative and quantitative models are
aggregated to assist with the definition of a “to-be” system. Thereafter, mappings of cloud environments are made to both the “as-is” and “to-be”
systems so that functional and performance requirements are defined for cloud applications.

A 3.0 Consolidation Via Virtualization

Virtualization enables the abstraction and aggregation of data center resources in order to transition them into a unified logical structure that can
be shared by all application loads. Virtualization decouples the physical IT structure from hosted services.
Consolidation is a critical application of virtualization by enabling IT departments to gain control of distributed resources.
Virtualization of services, storage and networks not only allows the mobility of applications within a singular data center, but also across multiple
centers and networks.

A 4.0 Security

Security practices for clouds are state-of-the-art and are available for all facets of monitoring and control. Private clouds employ typical methods
now in-place for all large-scale IT systems. Public clouds are protected by the same processes as are used by interactive enterprise services.

A 5.0 Optimization of Access and Processing Usage

Commercial off-the-shelf (COTS) tools exists which assist with the optimal usage of all cloud resources. The tools also provide a capability to
expand the number of user interfaces with a network dynamically and to supervise operations so that large bandwidths are available for resource
sharing.

A 6.0 Collaborative Processing

In addition to the optimization of access and processing of applications programs, collaborative processing, when it exists, can also be addressed.
This is an extension of access optimization because timelines are modeled and reduced for multiple users that are exchanging data, and sharing
software and other resources with the intent of converging upon a solution to a complex set of computations.

A 7.0 Encapsulated Decision Support

An automated decision support system transitions domain data imbedded in a cloud to knowledge representation and actionable intelligence.