Why Architecture Reviews?
Architecture reviews create transparency in software development, identify risks and uncover trade-offs. They make problems quantifiable and effectively focus investment in the technical improvement of systems.
Architecture Reviews at a Glance
What?
We scale architecture reviews from a small quick check to a large evaluation. We draw on established review standards such as ATAM or DCAR as needed, use smaller methods such as the Pre-Mortem, or rely on our own lightweight approach (LASR).
When?
Reviews are advisable when big / direction-setting architecture decisions are pending, larger delivery difficulties exist, central quality goals risk being missed, investments or purchase decisions need to be secured, or communication in and between development teams isn’t working optimally.
Why?
Your benefit from embarc’s support lies in our methodical leadership in the field of software and architecture reviews and our independence. Together we uncover critical risks, create decision confidence, and, where needed, give recommendations for further development.
Approach
Architecture reviews can take a few hours, or last several days or weeks. In very small forms, we rely on expert assessments or our own method, LASR (Lightweight Approach for Software Reviews). In larger contexts, we craft a suitable and efficient approach from methods such as DCAR, ATAM and others. The process outlined here is always fundamentally similar, but is individually and precisely tailored to your situation. We also go through steps 02 to 04 iteratively for large evaluations.
00
Initial Clarification
We define a goal for the review and simultaneously set a time horizon for results. The review model we apply follows from these two things plus the context of the solution, domain, stakeholders etc.
01
Presenting the Architecture
To build an understanding of the solution, we need an early overview of the system’s architecture. What technologies, concepts and ideas are being used? Where does the system come from and where should it evolve to? We give initial feedback and plan follow-ups.
02
Detailing the Goals
For larger reviews, we detail the goal level further with the desired system properties. The result is an evaluation standard with concretely formulated quality scenarios - a “Utility Tree” in ATAM terms. For smaller reviews, rough statements about quality goals are enough for us.
03
Analyzing the Code
This step is especially relevant in reviews with strong maintainability goals. At the code level, the concrete use of technologies and frameworks can be inspected, inconsistencies (with the architecture idea) can be recognized, and hot spots identified.
04
Discussing the Details
In workshops or short meetings with a changing set of participants, we work our way deeper into the software solution. We derive risks and trade-offs and identify strengths or weaknesses that stand out to us. How detailed and structured these discussions are depends on the size of the review.
05
Preparing the Results
Depending on the need, we can turn the results of a review into recommendations, work directly with teams to eliminate risks, or prepare findings in a more abstract form. There are options ranging from a short task list to a management report and presentation.
Did we spark your interest?
Unsere Expertise
Our consultants have years of experience conducting architecture reviews across a wide range of industries. In total, we have already performed over 100 architecture evaluations. This knowledge has flowed into many publications and the book “Leichtgewichtige Software Reviews”. embarc also authored the standard curriculum of the iSAQB on this topic (ARCEVAL).
Our books, articles and blog posts on this topic
Book
Software-Systeme reviewen mit dem Lightweight Approach for Software Reviews - LASR (German)
More information
Article
Leichtgewichtige Software-Reviews mit LASR (German)
Informatik Aktuell, June 2024
Read the article
Topic page
Our overview on architecture reviews (German)
Topic page: Architekturreviews
Software-Systeme reviewen mit dem Lightweight Approach for Software Reviews - LASR (German)
More information
Article
Leichtgewichtige Software-Reviews mit LASR (German)
Informatik Aktuell, June 2024
Read the article
Topic page
Our overview on architecture reviews (German)
Topic page: Architekturreviews
FAQs
When are architecture reviews advisable?
Reviews of software or architecture involve several developers and other roles. Even small reviews should therefore have a concrete occasion. For very lean reviews (e.g. following LASR - Lightweight Approach for Software Reviews), a communication deficit between teams can already be enough of a reason. For larger reviews, there should be clear uncertainties in the architecture area, larger investments pending, or production problems present.
What's the outcome of a review?
Every review names risks or problem areas of the examined solution. So you at least know which partial solutions are (potentially) poor. The more elaborate the approach, the more likely strengths and weaknesses are also classified, trade-offs in the solution design identified, goal achievement assessed in more detail, or recommendations for further development given. The form ranges from simple lists of findings to management-compatible evaluation reports.
Are reviews "from the outside" really useful?
Being evaluated by external partners has advantages and disadvantages. External partners have less domain knowledge and, above all, less solution knowledge. This makes evaluations somewhat more elaborate and they may stay a bit more on the surface. On the other hand, the outside perspective also enables fresh impulses, an independent view, and bubble-free insights. The methodical competence of external reviewers can also help. This pays off most in medium to large evaluations.