Introduction

This appendix contains a list of criteria to be used in performing a testing process health check and a suggested approach to conducting the task.

The testing process health check could be used in a number of ways:

• To baseline the current state of testing within an organization prior to introducing a formal testing process

• To monitor the progress of the adoption of a formal testing process, and to identify areas that were satisfactory and areas capable of improvement

• As part of a process improvement program (see Chapter 13 – Improving the Testing Process).

The criteria used in the testing process health check are grouped under the following high-level headings:

• Testing and the software development process

• Testing process

• Roles and responsibilities

• Testing phases.

These groups and their individual criteria are described in full in the following sections respectively. Section Q7 contains an optional scheme for scoring the various criteria if a more quantitative approach is required.

A checklist is provided at the end of this appendix that can be copied and used as an aide-memoir during the testing process health check. This checklist is also available to download using the following link: http://www.cambridge.org/9780521148016/.

Terminology

The term “subject” is used to denote the person, project, department, or organization that is the target of the testing process health check. The term subject could refer to a third-party organization whose testing process needs to be assessed, or internally within the host organization.

Part 3 of this book contains a set of standard testing document templates, proformas, and checklists, and a number of appendices that expand on topics described in passing in the main body of the book.

The standard testing document templates, proformas, and checklists are also available from the following link: http://www.cambridge.org/9780521148016, so that they can be used immediately without modification or customized to reflect the particular requirements of any organization (such as a corporate style, branding, or documentation standard).

Introduction

This appendix contains an example illustrating the testing technique of equivalence partitioning described in Chapter 3, which the test analyst can use to select specimen data for use in testing the application under test (AUT).

The Testing Problem

The specification for a software system for validating expenses claims for hotel accommodation includes the following requirements:

• There is an upper limit of £90 for accommodation expense claims

• Any claims above £90 should be rejected and cause an error message to be displayed

• All expense amounts should be greater than £0 and an error message should be displayed if this is not the case.

Analyzing the Testing Requirements

To support the process of analyzing the previous requirement, it is useful to graphically show the partitions and their boundaries and to state the ranges of the partitions with respect to the boundary values (see Figure K1).

Designing the Test Cases

The next step is to design the test cases by drawing up a table showing the test case ID, a typical value drawn from the partition to be input in the test case, the partition it tests, and the expected output or result of the test case (see Table K1).

Purpose of the Book

This book provides comprehensive and rigorous guidance to workers in the field of software testing for researching or setting up a software testing process within organizations.

The book provides advice and guidance on all aspects of the testing process, including:

• The need to test software and the approach to testing

• Specific details of testing techniques, with examples

• The planning and management of testing projects

• Testing roles and responsibilities

• Comprehensive details of the testing phases

• Extensive testing document templates, proformas, and checklists

• Recommendations for testing process improvement and the role and use of metrics

• The testing challenges facing testers involved in quality assurance tasks on agile projects

• The testing challenges facing developers of object-oriented and component-based systems

The book covers the testing of software from a number of sources, including software developed or modified in-house, software that represents the modification or extension of existing legacy software systems, and software developed on behalf of an organization by a third party.

The book also covers acceptance testing of commercial off-the-shelf (COTS) software procured by an organization, or COTS software that has undergone bespoke development either internally or by a third party on behalf of an organization.

A few technologies have been crucial in enabling the development and use of cloud platforms. Web services allow applications to communicate easily over the internet: Composite applications are easily assembled from distributed web-based components using ‘mashups.’ If there is one technology that has contributed the most to cloud computing, it is virtualization. By decoupling the software platform from hardware resources, virtualization enables massive cloud data centers to function seamlessly in a fault-tolerant manner. Similarly, multi-tenancy allows the same software platform to be shared by multiple applications, and can thus be looked upon as application-level virtualization. Multi-tenancy is critical for developing software-as-a-service applications and Dev 2.0 platforms.

In this part of the book we review the fundamentals of enterprise software architecture: The information needs of many large enterprises are similar, a fact that has been exploited by packaged ERP, CRM and SCM products. Understanding how enterprise information is modeled assists in making choices regarding the applications an enterprise needs, i.e., whether to buy packaged products or build custom solutions, as well as whether to deploy systems in house or in the cloud. Software architecture principles are the key to building custom enterprise applications efficiently. Abstracting architecture patterns leads to re-use, and eventually to Dev 2.0 platforms that enable simple applications to be built without any programming. Modeling business processes and executing them using workflow engines is also a mature example of how abstraction can lead to re-use. Finally, decision support and knowledge management are becoming increasingly crucial functions of enterprise applications. Therefore, we present an overview of search and analytics techniques using a unified, and hopefully illuminating, matrix algebra formulation.

So far we have covered a number of cloud computing technologies as well as explored their impact on the software needs of enterprises. In the process we have limited our discussion to the major cloud computing providers, viz. Amazon, Google and Microsoft, with most of our examples taken from the first two, given that Microsoft's offering is still in its nascent stages at the time of writing.

However, the cloud computing ecosystem includes other cloud providers, albeit smaller than the three major ones. Additionally, there are a range of emerging technologies that complement public clouds, enable interoperability between private data centers and public clouds, or facilitate the creation of private clouds within enterprises.

Figure 17.1 depicts our classification of the cloud ecosystem from an enterprise perspective, also indicating the organizations involved in creating and bringing these technologies to market. Needless to say this list of organizations is incomplete and evolving; moreover, given the rate of innovation in the cloud space, it is possible that additional technology categories may emerge in the future. (Note: there are other similar classifications, such as the OpenCrowd taxonomy, which includes a far broader range of technologies and applications.)

The enterprise cloud ecosystem comprises of three main categories; cloud service providers, tools for using and managing cloud deployments and tools for building private clouds.

Barely 50 years after the birth of enterprise computing, cloud computing promises to transform computing into a utility delivered over the internet. A historical perspective is instructive in order to properly evaluate the impact of cloud computing, as well as learn the right lessons from the past. We first trace the history of enterprise computing from the early mainframes, to client-server computing and 3-tier architectures. Next we examine how the internet evolved into a computing platform for enterprise applications, naturally leading to Software as a Service and culminating (so far) in what we are now calling cloud computing. Finally we describe how the ‘enterprise architecture’ function within IT departments has evolved over time, playing a critical role in managing transitions to new technologies, such as cloud computing.

The internet is based on a universally accepted set of protocols, HTTP, DNS, and TCP/IP, that provide the foundation for web-based cloud computing offerings. In this chapter we examine three critical web-based technologies at the next level of granularity that have been instrumental in improving the usability of web-based applications: Web services are used to request for and access infrastructure services in the cloud; AJAX-based user interfaces allow web-based applications to be user friendly; finally mashups bring a new dimension to software as a service by enabling users to compose multiple SaaS applications into a single user interface.

WEB SERVICES: SOAP AND REST

We have discussed the genesis and evolution of web services briefly in Chapter 2. Here we give a brief technical overview of both SOAP/WSDL and REST-based web services, and also compare these in the context of their utility in building cloud-based offerings. For a more detailed description of web services protocols, see [30].

SOAP/WSDL Web services

SOAP/WSDL web services evolved from the need to programmatically interconnect web-based applications. As a result SOAP/WSDL web services are essentially a form of remote procedure calls over HTTP, while also including support for nested structures (objects) in a manner similar to earlier extensions of RPC, such as CORBA; we shall return to this point later.

The elements of a SOAP/WSDL web service are illustrated in Figure 7.1, using as an example the service provided by Google for searching the web.

The emergence of cloud platforms has given rise to new paradigms for dealing with distributed data in the cloud, parallel computing using very large computing clusters as well as rapid application development tools for specialized domains: Cloud-based data stores differ significantly from traditional relational databases, with different query and consistency semantics as well as performance behavior. The MapReduce programming paradigm makes large-scale analytics tasks easy to define. MapReduce implementations allow massive computing clusters to be used while tolerating faults that are inevitable at such scales. Similarly, but in a very different context, Dev 2.0 platforms allow simple business applications to be developed by end-users using always-on hosted platforms in the cloud, obviating the need for traditional development and thereby increasing business agility.

INTRODUCTION

By ‘enterprise computing’ we mean the use of computers for data processing in large organizations, also referred to as ‘information systems’ (IS), or even ‘information technology’ (IT) in general. The use of computers for enterprise data processing began in the 60s with the early mainframe computers. Over the years enterprise computing paradigms have changed dramatically with the emergence of new technology: The advent of the PC in the 80s led to the replacement of large mainframe computers by client-server systems. The rise of the internet in the 90s saw the client-server model give way to web-based enterprise applications and customer-facing e-commerce platforms.

With each of these advances, enterprise systems have dramatically improved in terms of scale and ubiquity of access. At the same time their complexity, and consequently cost, has increased as well: Trillions of dollars are spent world-wide on information technology, including hardware and software purchases as well as application development (in-house or outsourced). It is also estimated that enterprises spend between two and ten percent of their revenues on IT.

Now, cloud computing offers the potential for revolutionizing enterprise computing once more, this time by transforming computing itself into a utility that can be accessed over the internet. In his recent book The Big Switch, Nicholas Carr compares the possible ramifications of such a change to the creation of the electricity grid in the early twentieth century.

Applications have traditionally been developed for use by a single enterprise; similarly enterprise software products are also developed in a manner as to be independently deployed in the data center of each customer. The data created and accessed by such applications usually belongs to one organization. As we discussed earlier in Chapter 3, hosted SaaS platforms require a single application code to run on data of multiple customers, or ‘tenants’; such behavior is referred to as multi-tenancy. In this chapter we examine different ways to achieve multi-tenancy in application software.

Before proceeding it is important to also note that virtualization, as discussed in the previous chapter, is also a mechanism to achieve multi-tenancy at the system level. In a virtualized environment, each ‘tenant’ could be assigned its own set of virtual machines. Here we examine alternatives for implementing multi-tenancy through application software architecture rather than at the system level using virtual machines. Thus, such multi-tenancy can also be termed application-level virtualization. Multi-tenancy and virtualization are both two sides of the same coin; the aim being to share resources while isolating users from each other: hardware resources in the case of system-level virtualization and software platforms in the case of multi-tenancy.

MULTI-ENTITY SUPPORT

Long before ASPs and SaaS, large globally distributed organizations often needed their applications to support multiple organizational units, or ‘entities,’ in a segregated manner.

As is well known, the internet was born as a communication infrastructure for data sharing between large government research labs in the US, and soon grew to include academic institutions across the world. The development of the NCSA Mosaic web-browser in 1993 sparked the rapid expansion of internet use beyond these boundaries into a platform for sharing documents, using the HTTP protocol and HTML markup languages developed by Tim Berners Lee at CERN, Geneva, in 1990. Using a browser, information ‘published’ over the internet could be accessed anonymously by the public at large, giving rise to the ‘world wide web’. The subsequent history of the commercialization of the web and the dot-com boom is also well known. In this chapter we explore how and why the internet also evolved into a platform for enterprise applications, eventually giving birth to the cloud computing paradigm.

INTERNET TECHNOLOGY AND WEB-ENABLED APPLICATIONS

Internet-based applications rely fundamentally on HTTP, the HyperText Transfer Protocol, and HTML, the HyperText Markup Language; both are now standards defined by the world wide web consortium (W3C). Browsers, such as Internet Explorer, and servers, such as HTTPD (HyperText Transfer Protocol Daemon) implement these standards to enable content publishing over the internet. Other technologies such as XML and SOAP are also important, and will be covered in later chapters. Here we review the essential aspects of these underlying technologies that are critical to understanding internet-based enterprise applications and cloud computing.

We are nearing the conclusion of our exploration of cloud computing; along the way we have also covered many aspects of enterprise architecture. Certain aspects of enterprise architecture may need to adapt to or be replaced by emerging paradigms, such as PaaS or Dev 2.0. At the same time many enterprise applications can, technically, be easily deployed in an IaaS cloud. We have also examined the question of cloud economics and shown that in principle public cloud can be cheaper, due to elasticity, and also faster to use, due to automated provisioning. We have also noted the additional cost advantages of PaaS platforms that enable public facing applications to be made available at no cost, with charges accruing only once load increases.

So, where could, and should, an enterprise begin leveraging the cloud? There are valid concerns with using public clouds from an enterprise perspective, especially with respect to (a) data confidentiality, lock-in, and auditability as well as (b) software licensing. Enterprises are still wary of placing production or sensitive data in the cloud, since current cloud offerings are essentially public networks and hence exposed to more attacks. While there are no fundamental obstacles to making a cloud environment as secure as an in-house data center, this requires careful planning using encrypted storage, virtual LANs and network middleware. Some cloud providers have begun to take the first steps towards these levels of security, such as Amazon's VPC.

In the previous chapter we saw how basic forms-based transactions are assembled into an application. The resulting application resembles a ‘playground’ where each user can access available forms to create, modify or view data. In the process any business logic attached to the forms is also executed. In such a scenario, it is up to the user to decide what form to access and which transaction to perform. However, in real life users do not perform such tasks in a vacuum; rather they perform these tasks as part of an overall business process in which the user is participating. In this chapter we examine mechanisms by which an enterprise application can track such business processes and drive the ‘flow of work’ to users.

IMPLEMENTING WORKFLOW IN AN APPLICATION

We consider a simple ‘leave approval’ process. An employee places a leave request that goes to a supervisor. If approved it is forwarded to the HR department for their records; simultaneously the result, i.e. approval or rejection, is conveyed back to the employee. First let us design an application that implements this process, using a data model as shown in Figure 15.1. A LEAVEREQUEST table keeps track of the requests placed. The EMPLOYEE table captures each employee's supervisor as a self-referential foreign key. The application provides a New-Leave-Request form using which an employee can apply for leave.