Raising the level of cloud design, ForkMe!

When preparing a design, project owners and planners are not always able to accurately dictate the requirements for using the cloud. If vague requirements such as "normal cost" or "good performance" are left alone, there is a risk of cost explosion or system downtime due to high load, leading to litigation issues. In addition, presenting assumptions, such as the basis for the estimate and the scope of processing, is labor-intensive. If the parties involved do not have a firm grasp of these assumptions, the risk of misunderstandings among them being exposed when a problem arises, leading to trouble over where the responsibility lies.

When exchanging information with stakeholders for requirement confirmation and review, it is difficult to anticipate the necessary information from the outset and optimize the materials. On the other hand, if you prepare documents as needed, you will waste time preparing multiple documents with similar information, or you will have to interrupt your work to prepare supplemental materials. If an engineer with high design skills loses productivity by working on documents, it is a great loss for both the engineer and the team.

When preparing design documents, it is difficult to take into account the diversion of information for other projects. Therefore, the information is usually organized as project-specific information and stored in the back of a folder. The index of information across projects remains only in the minds of a limited number of people in charge. It is a loss to throw away valuable information that is worth reusing, such as how much performance should be assumed for similar campaigns.

To eliminate lack of confirmation, it is effective to adjust interview items to be used in the preparation stage and adjust precondition figures to be used in the review stage. For interviews, by following predetermined questions, it is possible to avoid omitting requirements that depend on the skills of the person in charge. Furthermore, if the items are adjusted to a granularity that makes it easy for the project owner or planner to answer or for the designer to make a Fermi estimation, it will be possible to make a certain level of inferences even if there are uncertainties. In addition, the use of predetermined numerical indicators for the presentation of assumptions avoids inadequate explanations that depend on the skills of the person in charge of the project. Furthermore, if they are measurable numbers, If they are measurable, it will be possible to isolate whether or not they are within the scope of design commitments for which responsibility should be borne when problems occur.

An effective way to reduce the amount of unnecessary materials is to adjust the document format. By creating documents according to a predetermined format, there is no need for additional materials due to insufficient descriptions dependent on the skills of the person in charge, and no need to expend unnecessary labor due to excessive descriptions. Moreover, if the format is adjusted to include information that is common to various applications, it avoids the waste of creating similar documents for different purposes.

In order to utilize valuable design information for other designs without discarding it, it is effective to adjust the data format in which the information is recorded. By avoiding variations in data format among projects and personnel, information can be easily retrieved and reused across projects. In addition, if the format is code such as JSON/YAML, it is possible to suppress dependence on specific tools and fluctuations in expression caused by charts and diagrams, thereby avoiding the burying of information.