What is a TOP Infrastructure

If you are a CTO, what does a top infrastructure mean to you? The following points might be one part of your answers. These points are not Utopie. We will provide a framework to achieve these objects in the following posts. Our SAP hosting service: saps.top-infra.com is an implementation example.

Highest availability

Nothing is more important than the system availability, but unfortunately, a real 100% availability system only exist in commercial presentation. The system must provide end user with the highest availability within the given budget. It is also interesting to know that the downtime might be unperceived by the end users. It could be a few seconds for e-commercial site or a few minutes for some transactional system. It is the human nature to verify and retry before concluding that the system is really down, when the system becomes alive before the retry, no downtime is perceived by the end user. It might be a perfect answer to the question of high availability.

Good performance assured

A perfect system should have no performance issue in any excuse. You might prefer to hear “We have spent much more money in CPU power because there are much more user activities” then “Our system is quite slow this month because there are more concurrent users than expected.” When system resource is elastic with the resource requirements, the good performance could be assured.

Lowest infrastructure cost

What is the lowest infrastructure cost? There is a theoretical answer: Pay-Per-Use. Attention: it is not like the IaaS such as the AWS cloud service. AWS is good but not great: a dedicated 4 core CPU / 16 GB RAM server costs about 45 $ per month while the equivalent AWS m4.xlarge instance costs 185$ per month. The theoretical lowest pay-per-use cost should be based on a dedicated infrastructure. In the given example, when you use AWS, you pay about 400% of the theoretical lowest cost. We should do it much better. The lowest infrastructure cost also means reduce of the license cost and the maintenance taskforce.

Minimized no-production systems cost

It might be difficult to explain to business why they must pay for a preproduction system, a quality system and a development system. Even the business accepts the budget, we all knows that the company’s central interest should focus on the production system. The total cost of no-production systems should be minimized.

Least human tasks

Human tasks are not only expensive but also a source of errors. In a well-designed landscape, all of the most repetitive tasks should have already been automated.

Well-designed architecture

As is said “In the beginner’s mind there are many possibilities, in the expert’s mind there are few.” A well-designed architecture can help to reduce the hardware requirement, license cost and maintenance taskforce. And that is purpose of our service.

Posted in Best Practices.

Leave a Reply

Your email address will not be published. Required fields are marked *