{"id":3234,"date":"2026-08-28T08:44:00","date_gmt":"2026-08-28T08:44:00","guid":{"rendered":"https:\/\/www.torck.io\/?p=3234"},"modified":"2026-10-30T15:48:39","modified_gmt":"2026-10-30T15:48:39","slug":"modernise-legacy-software","status":"publish","type":"post","link":"https:\/\/www.torck.io\/en\/legacy-software-modernisieren\/","title":{"rendered":"Modernize or Replace Legacy Software? A Decision Tree"},"content":{"rendered":"<p><img fetchpriority=\"high\" decoding=\"async\" src=\"https:\/\/www.torck.io\/wp-content\/uploads\/2026\/08\/blog-header-legacy-software-modernisieren.svg\" alt=\"Cover Image: Decision Tree for Legacy Software: Modernize or Replace\" width=\"1600\" height=\"900\" loading=\"eager\" \/><\/p>\n<p>Modernizing legacy software or replacing the legacy system entirely: This is rarely a purely technical question. Business value, available resources, and the actual condition of the system all play a role. Many companies, out of habit, continue to use the old system, or, out of impatience, opt for something completely new without seriously considering the options in between.<\/p>\n<div class=\"tk-key\"><b>Short answer<\/b><\/p>\n<p>Anyone looking to modernize legacy software has four intermediate options between continuing to use the existing system and building a new one. The <a href=\"#sechs-r\">6R model<\/a> categorizes them as follows: Retire, Rehost, Revise, Refactor, Rebuild, Replace. Four factors determine which option fits: technical condition, business value, available resources and strategic goal. A stable but hard-to-maintain system usually needs Refactor, not Replace.<\/p>\n<\/div>\n<h2 id=\"sechs-r\">Modernizing Legacy Software Using the 6R Model<\/h2>\n<p>The 6R model distinguishes six approaches for legacy systems, ranging from the smallest to the largest change (<a href=\"https:\/\/ardura.consulting\/de\/blog\/modernisierung-von-altsystemen-wann-neu-aufbauen-refaktorisieren-oder-ersetzen\/\" target=\"_blank\" rel=\"noopener\">ARDURA Consulting<\/a>, 06\/2025).<\/p>\n<div class=\"tk-tablewrap\">\n<table>\n<caption>An overview of the six options, according to ARDURA 06\/2025<\/caption>\n<thead>\n<tr>\n<th>Option<\/th>\n<th>What Happens<\/th>\n<th>This works if<\/th>\n<\/tr>\n<\/thead>\n<tbody>\n<tr>\n<td>Retire<\/td>\n<td>System will be shut down without a replacement<\/td>\n<td>There is no longer any measurable goodwill<\/td>\n<\/tr>\n<tr>\n<td>Rehost<\/td>\n<td>Move to new infrastructure without changes<\/td>\n<td>It's just the platform that's outdated, not the code<\/td>\n<\/tr>\n<tr>\n<td>Revise<\/td>\n<td>Minor technical adjustments without architectural changes<\/td>\n<td>A few weaknesses are bothersome, but not the whole picture<\/td>\n<\/tr>\n<tr>\n<td>Refactor<\/td>\n<td>Rewrite the code structurally; the functionality remains the same<\/td>\n<td>The code is stable but difficult to maintain<\/td>\n<\/tr>\n<tr>\n<td>Rebuild<\/td>\n<td>New development based on the existing logic<\/td>\n<td>The architecture is no longer expandable<\/td>\n<\/tr>\n<tr>\n<td>Replace<\/td>\n<td>Replacement with a completely new solution<\/td>\n<td>Technology and business logic are both at an end<\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n<\/div>\n<h2 id=\"entscheidungsbaum\">The Decision Tree in Seven Questions<\/h2>\n<p>The order of the questions determines how quickly an option can be ruled out.<\/p>\n<ol>\n<li><strong>Does the system still provide measurable business value?<\/strong> No, then retire it. This question is often resolved more quickly than expected when no one really knows the usage figures.<\/li>\n<li><strong>Is the underlying technology a total loss?<\/strong> If the manufacturer is no longer around, there is no longer any support, and there is no developer community, then there is no alternative to rebuilding or replacing the system.<\/li>\n<li><strong>Is the architecture fundamentally extensible?<\/strong> If not, even if the code works, this limits all modernization options except for a rebuild or a replace.<\/li>\n<li><strong>Is the code technically stable but difficult to maintain?<\/strong> This is exactly where Refactor fits in best (ARDURA, 06\/2025). The business value is there; the technical problem lies in the code quality, not in the architecture.<\/li>\n<li><strong>Is a new infrastructure enough without changing the code?<\/strong> In that case, rehosting is the fastest and most affordable option.<\/li>\n<li><strong>Are there only minor technical issues?<\/strong> In that case, Revise is sufficient, without any major changes.<\/li>\n<li><strong>Is there no clear answer to any of the questions one through six?<\/strong> In that case, a more in-depth technical assessment is needed before a decision is made. Hasty decisions at this stage are the most common reason why modernization projects fail later on.<\/li>\n<\/ol>\n<h2 id=\"was-entscheidet\">What Factors Determine the Choice Between Modernizing and Replacing<\/h2>\n<p>According to ARDURA, choosing the right option depends on four factors: technical condition, business value, available resources, and strategic goals (ARDURA, 06\/2025). A system with high business value and poor technical condition justifies investment. A system with low business value rarely justifies investment, regardless of its technical condition.<\/p>\n<p>In practice, those looking to modernize legacy software often skip refactoring because \u201creplace\u201d seems like the more thorough approach, a new system free of old baggage. This is a fallacy if the legacy system has high business value and the code is merely difficult to maintain but technically stable. In this case, a rebuild or replacement costs significantly more time and money than the actual problem, maintainability, justifies.<\/p>\n<div class=\"tk-inline-cta\">\n<p><b>Assess Your Legacy System Based on the Seven Questions<\/b>We'll review your architecture, dependencies, and usage metrics and let you know which of the six options is the best fit. Initial consultation: 30 minutes, free of charge, with a response within one business day.<\/p>\n<p><a class=\"tk-btn\" href=\"\/en\/?anfrage=legacy-software-modernisieren#contact-us\">Request a Review<\/a><\/div>\n<h2 id=\"laufender-betrieb\">Modernization During Operation<\/h2>\n<p>Rehost and Revise are usually carried out alongside day-to-day operations. Rebuild and Replace, on the other hand, almost always require a transition phase with parallel operation, during which the old and new systems run simultaneously.<\/p>\n<div class=\"tk-note\"><b>The item that is most often missing<\/b>The parallel operation phase is rarely short and rarely cost-effective. It should be included in the planning from the very beginning, not as a retroactive adjustment once both systems are already incurring costs simultaneously.<\/div>\n<p>One contractual detail is often overlooked in this context. Anyone replacing a legacy system should assess the extent to which the new provider or development partner creates a new lock-in. We describe the most important safeguards for this in the <a href=\"\/en\/avoid-vendor-lock-in\/\">Contribution to Avoiding Vendor Lock-in<\/a>. The breakdown of the costs over five years is provided in the <a href=\"\/en\/hidden-costs-software-development\/\">article on hidden costs<\/a>.<\/p>\n<h2 id=\"ohne-unterbrechung\">Replacing the legacy system without disrupting operations<\/h2>\n<p>It is rarely possible to shut down a legacy system on a specific date without disrupting day-to-day operations. For this reason, legacy systems are usually replaced in stages rather than in a single, major transition. Three components work together in this process.<\/p>\n<ul>\n<li><strong>Strangler Fig Pattern:<\/strong> The new system takes over one function at a time. An upstream layer, such as an API gateway or proxy, routes requests to either the old or the new system, depending on the function. The legacy system gradually shrinks until it no longer has any tasks to perform and can be shut down.<\/li>\n<li><strong>Parallel operation with reconciliation:<\/strong> For a transitional period, the old and new systems will process the same transactions. Discrepancies in the results will become apparent before users begin working exclusively with the new system. A defined fallback plan to the old system is part of the planning.<\/li>\n<li><strong>Data Migration in Steps:<\/strong> The data is not transferred in a single export, but rather by department or data class. Each batch is checked before the next one follows. As long as both systems are running, it must be clearly defined for each data type which system is the primary one.<\/li>\n<\/ul>\n<p>The right order depends on the dependencies in the legacy system. How we plan and support this kind of replacement is described on the page on <a href=\"\/en\/legacy-modernization\/\">Legacy Modernization by torck<\/a>.<\/p>\n<h2 id=\"haeufige-fragen\">Frequently asked questions<\/h2>\n<h3 class=\"tk-faq-item\" id=\"totalschaden\">How can you tell if a system is a total technological loss?<\/h3>\n<p>Typical signs include a manufacturer that no longer offers support, a programming language or platform for which it is difficult to find developers, and an architecture that no longer allows for new features without compromising existing ones. If several of these points apply, there is hardly any way around rebuilding or replacing the system (ARDURA, 06\/2025).<\/p>\n<h3 class=\"tk-faq-item\" id=\"im-betrieb\">Is it possible to modernize while operations are ongoing?<\/h3>\n<p>For \"Rehost\" and \"Revise,\" yes, because infrastructure changes or minor adjustments can usually be implemented without downtime. For \"Rebuild\" and \"Replace,\" parallel operation over weeks or months is the norm, requiring a corresponding investment of time and resources for both systems simultaneously.<\/p>\n<h3 class=\"tk-faq-item\" id=\"kosten-nichtstun\">How much does it cost to simply keep a legacy system running?<\/h3>\n<p>The costs of inaction are rarely immediately apparent, but they grow with each passing year. These include rising maintenance costs, increasing security risks, and a loss of adaptability to new business requirements. Before companies modernize their legacy software, it\u2019s worth assessing their own systems, as general figures are of little help here.<\/p>\n<h2 id=\"naechster-schritt\">The next step<\/h2>\n<p>torck evaluates legacy systems using the 6R model before making a recommendation to rebuild, refactor, or replace them. Our development teams in Maxh\u00fctte-Haidhof, Vienna, and Rabat are familiar with the pitfalls from their own modernization projects for industry and commerce. The contractual partner for this project is the German company torck GmbH. You can find out how we proceed on the page for <a href=\"\/en\/legacy-modernization\/\">Legacy Modernization<\/a>. During the initial consultation, we'll walk through your system together.<\/p>","protected":false},"excerpt":{"rendered":"<p>A decision tree based on the 6R model helps to objectively evaluate legacy systems, rather than continuing to operate them out of habit or replacing them out of impatience.<\/p>","protected":false},"author":10,"featured_media":3682,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"_mbp_gutenberg_autopost":false,"footnotes":""},"categories":[21],"tags":[],"art":[29],"class_list":["post-3234","post","type-post","status-publish","format-standard","has-post-thumbnail","hentry","category-individuelle-softwareentwicklung","art-entscheidung"],"_links":{"self":[{"href":"https:\/\/www.torck.io\/en\/wp-json\/wp\/v2\/posts\/3234","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/www.torck.io\/en\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/www.torck.io\/en\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/www.torck.io\/en\/wp-json\/wp\/v2\/users\/10"}],"replies":[{"embeddable":true,"href":"https:\/\/www.torck.io\/en\/wp-json\/wp\/v2\/comments?post=3234"}],"version-history":[{"count":9,"href":"https:\/\/www.torck.io\/en\/wp-json\/wp\/v2\/posts\/3234\/revisions"}],"predecessor-version":[{"id":7056,"href":"https:\/\/www.torck.io\/en\/wp-json\/wp\/v2\/posts\/3234\/revisions\/7056"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/www.torck.io\/en\/wp-json\/wp\/v2\/media\/3682"}],"wp:attachment":[{"href":"https:\/\/www.torck.io\/en\/wp-json\/wp\/v2\/media?parent=3234"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/www.torck.io\/en\/wp-json\/wp\/v2\/categories?post=3234"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/www.torck.io\/en\/wp-json\/wp\/v2\/tags?post=3234"},{"taxonomy":"art","embeddable":true,"href":"https:\/\/www.torck.io\/en\/wp-json\/wp\/v2\/art?post=3234"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}