Non functional Attributes (Quality Attributes)

• infrastructure services can be well defined as a function, like providing disk space, or routing network messages.

  • Functional Attributes

• Non-functional attributes, on the other hand, describe the qualitative behavior of a system, rather than specific functionalities.

• Examples of non-functional attributes are:

  • Availability

  • Scalability

  • Reliability

  • Stability

  • Testability

  • Recoverability

• NFAs important for the successful implementation and use of an IT infrastructure,

  • but in projects, they rarely get the same attention as the functional services

• When NFAs are not meant they have a direct impact on the business

  • Slow to deploy features to customers

  • Cannot maintain and users experience more bugs

  • Regulations cannot be met, and fines and reputational damage occurs

  • User experience is slow, buggy or outdated, and users will leave to competitors (clients or employees)

• NFAs are functional indeed, but they are not directly related to the primary functionalities of a system

  • Better term would be quality attributes

  • This leads to important stakeholders of a business (ones with money and influence) to not have the same feelings about NFAs and focus more on functionality.

    • They conisder NFAs as hygiene factor and taken for granted

    • They dont mention them, or state them explicitly

    • But they have expectations of them

Non functional Requirements (NFRs)

• Architects and engineers job is to find out these NFRs

• the acceptance of a system is largely dependent on the implemented non-functional requirements.

  • A website can be very beautiful and functional, but if loading the site (performance, a non-functional requirement) takes 30 seconds, most customers are gone!

• A large part of the budget for building an infrastructure is usually spent in fulfilling non-functional requirements that are not always clearly defined

• Infrastructure delivers manyof the NFRs

  • Examples

    • An application using an IT infrastructure built with several single points of failure will probably not reach very high availability figures, no matter how well the application is built

    • when the IT infrastructure is not designed to be scalable, the applications built upon it cannot introduce scalability as an afterthought. Will need to redesign and reimplement a new infrastructure which can be costly

    • When an IT infrastructure is setup to be highly available, a badly designed application can make the end result highly unreliable

    • security flaws on the processes level can undo all

• not unusual to have conflicting non-functional requirements in a system

  • Example

    • security versus user friendliness.

      • Users expect highly secured systems, but really don’t want to be bothered by password changes, smart card authentication, and other annoying security measures.

    • performance and cost.

      • Getting a high-performance system usually means getting more and faster hardware, and using strict implementation rules (which takes time to do and dependent on needs high cost engineers)