Product constraints, in the information technology world, refer to the "ilities."13 To provide an example of different product constraints, the following six "ilities" are discussed in more detail below.
The proceeding examples are crucial to the success of the Internet application.
One business model that is evolving is sometimes referred to as the "choiceboard." This is where the consumer, through the Internet, clicks on the features and prices of what he or she wishes to purchase. The manufacturer provides an online configurator that allows the consumer to select what options he or she is willing to buy and at what price. This new business model requires additional requirements that extend beyond software. This change also affects both the sales and manufacturing processes for the company.
Hence, the Internet application involves changes in the normal business process. In order for the product to succeed, changes in other business areas, or business communities, are required. The community represents the other business units that are impacted by the Internet application.
These changes may not be known when the Internet-based application is being designed for its first iteration. The application must, therefore, be adaptable, meaning that the application must continually change to meet the demands of the customer and the market. It is important to document this requirement to remind the project development team during design reviews of what will occur with each iteration of the product.
Reliability refers to the expectations of the users as to the continual availability of the Internet application (including performance and fault tolerance requirements). Users expect applications to be available on a 24-hour-per-day basis. Reliability requirements speak to the mean time between failures. The Internet application runs on equipment that the users expect to have a minimal mean time between failures. This may require backups in hardware and networks. These, combined with the location requirements, provide valuable input to the network engineers when designing the network infrastructure.
Reliability needs may vary by function and location, so it is important to build a table that correlates the reliability by either "when" or "how" requirements. Tables 3.6 and 3.7 illustrate a means by which specific reliability and performance requirements can be captured and organized. It is important to identify a pattern as to what functions really require high availability and when. With a stock-trading application, for example, the day that stock options become due, volumes are highest. If this is for a specific type of stock, say, an energy stock, the prior clientele may likely reside in Texas. To guarantee reliability for that location on their biggest trading day, equipment may be diverted to support the one-day peak instead of having excess capacity for the entire year. In any case, the requirements need to be explicit so those in charge of planning the capacity for hardware and networks have sufficient knowledge of the users' expectations.
Table 3.6 Reliability (Mean Time Between Failures, MTBF) Requirements
Table 3.7 Reliability (Performance) Requirements
ity for that location on their biggest trading day, equipment may be diverted to support the one-day peak instead of having excess capacity for the entire year. In any case, the requirements need to be explicit so those in charge of planning the capacity for hardware and networks have sufficient knowledge of the users' expectations.
A good example of the importance of scalability was recently depicted in an advertisement in which a small group of entrepreneurially spirited individuals are standing around a personal computer. They are watching the opening of their new business venture, holding their breath as they eagerly await the first customer's order. Elated screams of success grow as the customer orders grow beyond their dreams. Then, quickly, the happiness turns to panic as the orders keep growing and growing. How are they going to satisfy what is now an unanticipated order volume?
This commercial illustrates the need for scalability on two fronts: technology and business community. Technology is the infrastructure that supports the Internet application. Business community is the business infrastructure that supports the product claims.
First, the technology. The technology must be scalable to adapt to the growth of the Internet application without the continual need to throw everything away and start over, the cost of which would deplete any revenue gain. Scalability is building an infrastructure that supports both vertical growth (increase in the number of sales for a specific product) as well as horizontal growth (increase in breadth of product).
It is important to anticipate some growth. Growth will impact the performance of the Internet application. The infrastructure must be flexible enough to support growth in network size, transaction handling, data throughput, page load timings, and security. This requires sophisticated tools that can model capacity planning on the site and network.
With regard to scalability technology requirements, it is important to document the potential growth in increments. A spreadsheet (see Table 3.8) or some CASE tools can be used to document these associative requirement details.
Table 3.8 Capacity Requirements Spreadsheet
What is important is to determine the minimum, average, and maximum number of occurrences for each requirement. In other words, it is important to determine the anticipated volume of events (when), the number of transactions (how), and the number of users (by each type/actor). You also need to capture the timeframe for these volumes. For example, some time-initiated events may be dormant except when the time is triggered (close of the stock exchange or the end of the business day for that country). It is important for the network engineers to model the performance under the worst-case scenarios.
Growth will also impact the business infrastructure that is put in place. It is important to anticipate the risk of a supplier not being able to meet the growth of your business. Some small companies went out of business when UPS went on strike! eCost took a beating when 3Com could not get the chips for Palm Pilots. Scalability, in a business sense, requires risk analysis of all dependent business partners.
B2B supply-chain applications involve multiple vendors using multiple vendors. Scalability business requirements need to anticipate the possibility of a supplier not being able to meet the necessary volumes. In most cases, you can split the anticipated volume across multiple vendors. The business has minimal control over what vendors your vendors use. The scalability business requirements must contain scenarios of supply-chain interruption.
First, identify the possible breaks in the supply chain. Anticipate the worst scenario as well as the probable scenario. Table 3.8 can be adjusted to determine the severity of each possible outage.
At the scope level, scalability will be defined as the need to implement a flexible infrastructure to support the projected growth. At this point, the requirement is still ambiguous; all the volumetrics will most likely not be known. This requirement evolves as more requirements are captured. It is still important to document the scalability need at this point as a reminder that more information is required.
At the planner level, volumetrics begin to emerge. The business units should be able to provide projected numbers of the customers, orders, transactions, and so forth. These volumetrics should be applied to all models. Relationships among functional requirements are also being formed at the business level. As part of defining the relationships, the business unit should be able to project volumetrics about the relationships. For example:
Each potential client will search for five product descriptions, including price and availability, at one time.
One out of three searching clients will, on average, order two products with each order placed.
One out of five customers will bypass the search and place an order for one item.
The bulk of orders are placed between 10 a.m. and 1 p.m. EST on Mondays. The exception is the seven weeks before the end of the calendar year. During that time, the peak period expands to all weekdays.
Large companies that are dabbling in the Internet must not take these volume estimates lightly. There is a possibility that the Internet-based product may outperform the corporation's standard business, making the Internet the leading business unit.
All of these business processes and information require some level of security processing. One type of security is required to protect the client, another to protect the corporation. The requirements engineer must spend time with all requirement suppliers on the impact of poor security. Security requirements must be captured on two fronts: (1) to protect the Internet user from intruders, and (2) to protect the corporation from theft or corruption (viruses).
It is critical to answer the simple question of what would be the impact to the corporation or the customer if data or processing were lost due to a security breach. If possible, attempt to obtain a quantifiable number at the following levels:
- Five-minute outage
- Day-long outage
- Week-long outage
Then try to determine through awareness-type questioning whether the data would actually be lost or corrupted. Sample scenarios include
- If data became corrupted
- If a competitor obtained any data
- If any file (data, voice, media) became corrupted
Clients must feel comfortable doing business with you. They must feel that you are reputable and that the payment process is a private transaction used for this purchase alone.
There are thieves among us. The first one who comes to mind is the predator who tries to destroy or capture corporate information.14 The other type of thief is one who wants to steal the product or service. When a thief of this sort obtains confidential information by using stolen payment options, it can cripple a company. Security must be taken very seriously and viewed as having a potentially negative impact at any level.
The seven worst security mistakes that senior executives make15 include four that can be documented during the requirements process:
Pretending the problem will go away if they ignore it
Relying primarily on a firewall
Failing to understand the relationship of information security to the business problem (understanding physical security but not seeing the consequences of poor information security)
Failing to realize how much money their information and organizational reputations are worth
It is up to the requirements engineer to capture security requirements and associate the need to "who," "what," "how," and "where" as well as any interactions between these functional focus areas.
The Internet has evolved current business models to a point where the consumer holds the power. This has changedand will continue to changethe way companies relate to their customers and compete with one another. Usability is a product constraint that directly speaks to the customer's ability to use the Internet product, and it must be captured. It should include
The maximum allowed response time
The availability to access 24 hours a day, seven days per week
The ability to customize views at specific points
The ability to filter information, minimizing what the user would perceive as "clutter data"
The availability of choices for the product
Most importantly, a consideration for usability is how many clicks it takes to retrieve information and how many clicks to complete a workflow. Response time can be eaten away with every screen of updated or new information. This translates to time that the user is not willing to give up if he perceives that the usability of the application is at fault. The requirements engineer should capture this type of need and associate the information with each "how" requirement, similar to how you related scalability requirements. This assists the designers (users of these requirements) in developing an effective product with an emphasis on usability.
Every one to three months (sometimes sooner), new features are offered in new product iterations. Many of these iterations are developed in parallel with each other. The work products, from the requirements to the code, must be easily maintained in order to keep to this rigorous schedule. Conforming to standards defined by the organization (and by international standards organizations) facilitates the making of products that are easily updated and maintained through each iteration. A general maintainability requirement must be defined at the owner level. As the relationships to other nonfunctional and functional focus areas are defined, the maintainability should be clarified as to its specific impact on the work products produced in those other focus areas. As the other requirements evolve, so should the maintainability requirement by describing the detail requirements as they relate to the structure and standard for the different perspectives and focus work products.