Glossary
IT system mapping
IT system mapping is the structured description of the components of an organization's information system and of the dependencies between them, from business process down to network device. It combines an inventory of those components with views that connect them, each view built to answer a specific question.
Updated
What is IT system mapping?
IT system mapping describes what an information system is made of and how its parts hold together: business processes, the applications that carry them, the servers and cloud services that host them, and the networks and filtering rules that connect them. It also goes by IT landscape mapping, IT mapping or IT infrastructure mapping.
A precise public definition comes from ANSSI, the French national cybersecurity agency, in its guide "Mapping the information system: How-to guide in 5 steps", published on November 21, 2018. Mapping must make it possible to draw up the inventory of the information system's assets, meaning the list of its components and their detailed description, and to present the system as views, which are partial pictures of its links and of how it operates.
That definition rules out two shortcuts. A diagram alone is not a map, because it says neither where its boxes came from nor what was left out. An inventory alone is not one either: a list of virtual machines does not tell you which one runs payroll. The word covers both the result and the work of maintaining it, which is why people look for a method as often as for an IT mapping tool.
What does a map of an IT estate contain?
The ANSSI guide splits a map into three visions that run from business to technology, each broken down into two views.
The business vision holds the ecosystem view (the entities and systems the information system interacts with) and the business view, which shows the system through its main processes and information. The application vision holds the application view (software components, the services they provide and the data flows between them) and the administration view, which lists the scopes and privilege levels of users and administrators. The infrastructure vision holds the logical infrastructure view (network partitioning, IP address ranges, VLANs, filtering and routing) and the physical infrastructure view, which describes the physical devices.
Each view is made of objects and attributes, listed in Appendix 1 of the guide. In each view a pivot object links to the adjacent views, which is what makes dependencies visible.
Our own reading as practitioners, not ANSSI's. The administration view is the one we see least often, yet it is the first thing people ask for when a privileged account is compromised. The application view nearly always exists, but without its flows; see our entry on application mapping. Flows and filtering usually live in a document of their own, the flow matrix, which the map should reuse instead of redrawing.
What is IT mapping used for?
The ANSSI guide ties mapping to four cybersecurity concerns: control of the information system, its protection (identifying the most critical and most exposed systems), its defense (responding to an incident and qualifying its impact) and its resilience (continuity planning and crisis management).
Responding to an incident or preparing an outage. A virtualization host has to be shut down for hardware maintenance: which virtual machines does it carry, which applications run on them, which business processes stop, who to tell? With a map, you read an impact analysis by following the links from layer to layer.
Handling end of support. Knowing that an operating system version is leaving support is not enough: you need to know which machines run it, which applications depend on them and who owns them. That link turns a list of versions into a plan for dealing with software obsolescence.
Preparing a migration. Changing hypervisor or leaving a data center requires knowing the flows of every machine. What gets forgotten is the quiet flow: a nightly backup, a monthly file transfer to accounting, a firewall rule opened for a contractor whose engagement has ended.
Passing an audit. The auditor picks a server and asks who owns it and which application it carries. The guide notes that diagrams are typically the first elements examined during audits and post-incident analyses.
Is IT system mapping required by NIS2 or ISO 27001?
The word "mapping" does not appear in Article 21 of Directive (EU) 2022/2555, known as NIS2. That article requires Member States to ensure that essential and important entities take "appropriate and proportionate technical, operational and organisational measures" to manage the risks to their network and information systems. Paragraph 2 lists ten areas the measures must at least include, among them "incident handling", "business continuity" and, under point (i), "asset management". Article 23 sets the deadlines: an early warning within 24 hours of becoming aware of a significant incident, an incident notification within 72 hours with an initial assessment of its severity and impact.
In practice you cannot manage assets you have not listed, and you cannot assess impact within 72 hours without knowing what depends on the affected machine. Mapping is a means of meeting these obligations, not an obligation written under that name in the directive.
One EU text is more explicit, for a limited scope. Commission Implementing Regulation (EU) 2024/2690 applies only to the digital entities listed in its Article 1, including DNS, cloud computing, data center and managed service providers. Point 12.4 of its annex requires them to "develop and maintain a complete, accurate, up-to-date and consistent inventory of their assets" and to record changes "in a traceable manner". For other entities the detail comes from national law; see our NIS2 entry.
Outside the EU, the NIST Cybersecurity Framework 2.0 points the same way: subcategory ID.AM-03 reads "Representations of the organization's authorized network communication and internal and external network data flows are maintained". On the ISO side, control 5.9 of ISO/IEC 27002:2022 is titled "Inventory of information and other associated assets", and the 27002 controls are carried into Annex A of ISO/IEC 27001, according to the journal of the ISO committee that writes both standards. The standards are paid documents and we do not paraphrase them.
How do you map an IT system? The five steps of the ANSSI guide
The guide's method has five steps.
Step 1, "How do I get a mapping project up and running?", identifies the stakes and the stakeholders and defines the scope, the target map and the path toward it. Step 2, "Which model should I adopt?", collects and analyzes what exists (inventories, diagrams, tools in place), then defines the model: objects, attributes, graphical representation and naming rules. Step 3, "Which tools should I use?", selects tooling able to compile the inventory, produce the views with their links and control the update process. Step 4, "How do I create my map step-by-step?", builds the inventory and then the views, incrementally (new views) and iteratively (refined views). Step 5, "How do I ensure my map remains relevant over time?", deals with communication and with keeping the map current.
The guide also proposes three maturity levels. At level 1 the map holds the core elements essential to cybersecurity operations, on a limited number of views. At level 2 all the views are present. At level 3 it is exhaustive and detailed. This is distinct from the level of detail assigned to each object and attribute, from 1 (essential information) to 3 (useful information).
Our comment. Step 2 is, in our view, the one teams rush most readily: the tool is opened before anyone has decided what counts as an "application" or how an environment is named. Step 4 is where the cost sits, because the inventory is completed through interviews and extractions.
IT mapping methods and tools compared
Six approaches coexist, and most IT departments combine several.
By hand, in a drawing tool or a spreadsheet. Visio, draw.io or a shared workbook is enough for a small scope, and nothing is faster for explaining an architecture to a committee. But a drawing has no model behind it: the same application carries three names on three diagrams, and updates depend on the author's memory.
The CMDB. The CMDB is the configuration database of IT operations: configuration items, relationships and owners, tied to tickets and changes. It describes the business layer and network flows less well, and it is only as good as the processes that populate it.
Enterprise architecture tools. They provide a repository with a metamodel, generated views and roadmaps, often in ArchiMate, which The Open Group describes as an open and independent modeling language for enterprise architecture. The ANSSI guide considers that dedicated software quickly becomes essential once the volume of data or the number of contributors grows. These tools are at their best on the business layer, the application layer and the target state. They need architects to populate them, and the technical layer, typed in or imported in batches, goes stale faster.
Open source. Mercator is a web application under the GPL license that manages the mapping of an information system as described in the ANSSI guide, whose data model it follows. Its documentation covers generated diagrams and reports, Excel import and export, and a REST API. It states one limit: for a large, complex and heterogeneous system, the model may prove incomplete.
Network discovery. Tools query the network and the machines (address scans, SNMP, agents, hypervisor interfaces) and report what answers. This approach finds what nobody declared. It sees neither the business layer, nor the owners, nor the reason a flow exists, and it needs accounts and network openings the CISO has to accept.
Mapping drawn from existing sources. The estate already describes itself in pieces: hypervisor export, CMDB extract, flow matrix, firewall rules, cloud accounts, architecture documents. This approach reads those sources, reconciles them and derives the maps, with no retyping. It keeps the provenance of every piece of information and shows where sources disagree. It knows nothing that no source contains, and it is as fresh as the last export.
How to choose an IT mapping tool
An IT mapping tool should be judged on what it looks like after a few months of use more than on its demo. Ten criteria follow; several come from the ANSSI guide.
Data intake comes first: how much is typed in, how much is imported or read, and what happens on the second import, when half the rows already exist? The model matters as much: does it adapt to your vocabulary without development, and does it cover cloud as well as on-premises sites? Layer coverage is checked on one case: from a business process down to a filtering rule, and back.
Provenance is the criterion demos show least. Does the tool say which source each piece of information came from, and on what date? When the CMDB and the hypervisor give two operating system versions for the same server, does it show the gap or silently pick one?
Four operational questions follow. Are the views generated from the inventory, or drawn next to it? Does the map export in a format readable by the main office software, as the guide strongly recommends? Can access be set view by view, since the guide advises restricting it? Where is the tool hosted? The guide asks that the map not be stored on the information system it depicts and that a backup stay accessible during a network failure.
That leaves simplicity, which the guide calls a major asset for avoiding "parallel" maps, and total cost: license, integration and above all the time of the people who maintain the data. Ask a customer of your size what keeping the map current takes.
Why IT maps go stale, and how to keep one current
A map is accurate on the day it is finished. The estate keeps changing: a machine created for a test and never deleted, a rule opened during an incident, an application migrated while its old server stays powered on. If updating is a separate task that sits in nobody's objectives, it loses to whatever is urgent.
The ANSSI guide devotes its fifth step to this. It names two good practices: regular update campaigns in which each stakeholder checks the information in their area, and an update stage built into projects that change the information system. It advises giving precedence to up-to-date macroscopic information over detailed information that is only updated sporadically.
We add one rule: whatever can be read back from a source should not be typed in again. The state of virtual machines is in the hypervisor, the state of rules is in the firewall, the state of cloud resources is in the accounts. A map regenerated from those sources at every collection only asks for human effort where it is useful: the link to the business, criticality, ownership, the reason for a flow. What is displayed still has to be dated: a map that says "based on the March 3 export" serves a crisis team better than an undated map everyone assumes is current.
Where to start an IT mapping project
With a scope you can finish. For a large information system, the ANSSI guide recommends starting with a map limited to certain views and centered on critical or exposed systems. Pick one question the map will have to answer (what happens if this cluster stops, which applications run on unsupported systems) and let it dictate the views.
Then gather what already exists, before any interview. The hypervisor export (an RVTools file for VMware, for example) lists virtual machines, their hosts, operating systems and networks; a separate article explains what to do with an RVTools export. The CMDB extract brings applications, owners and environments. The flow matrix says who should talk to whom, the firewall rules say who can, and the gap between the two is already a finding. Cloud accounts, read with a read-only role, complete the picture. Architecture documents, even outdated, tie applications to the business.
Decide early what identifies a machine (never its name alone, which repeats from one environment to the next) and what counts as an application. Then compare the sources before drawing anything. A server present in the hypervisor and missing from the CMDB, an application with no owner, a firewall rule no flow justifies: those gaps make up the first list of questions for the teams.
What Uncia brings, and what it does not do
Uncia applies the last of these approaches. It reads the sources of the IT estate in read-only mode, through connectors (clouds, hypervisors, CMDB, firewalls, scanners) and through files you drop in (exports such as RVTools, flow matrices, architecture documents), and reconciles them in a knowledge base. It generates Architecture, Infrastructure and Network maps in which you follow an element from business domain to firewall rule, each with a card that cites its sources. The same base gives the impact of an outage, end-of-support dates with their sources, and a check of architecture documents against the sources. Maps export to ArchiMate, draw.io, PNG and SVG. Uncia is designed to run in the customer's own environment; the language model proposes, a person decides. Details are on the IT estate mapping in Uncia page.
Its limits are those of the approach. Uncia changes nothing: it knows only what your sources contain, and the knowledge is updated at each collection, not in between. It complements your CMDB and your enterprise architecture tool and takes the place of neither: it reads your CMDB and your models and shows where they diverge from reality. The reports it structures along NIS2, ReCyF, DORA or ISO 27001 attach each section to the requirement it addresses, but a structured report is not a compliance verdict. And its three maps do not reproduce the six views of the ANSSI guide one for one.
Sources
- ANSSI, "Mapping the information system: How-to guide in 5 steps", published November 21, 2018 (presentation page), accessed
- ANSSI, text of the guide in English (PDF), accessed
- Directive (EU) 2022/2555 of December 14, 2022 (NIS2), Articles 21 and 23, EUR-Lex, accessed
- Commission Implementing Regulation (EU) 2024/2690 of October 17, 2024, Article 1 and Annex, point 12.4, EUR-Lex, accessed
- NIST Cybersecurity Framework (CSF) 2.0, NIST CSWP 29, Asset Management category (ID.AM), accessed
- ISO/IEC JTC 1/SC 27, SC 27 Journal, Volume 2, Issue 2, May 11, 2022, special issue on ISO/IEC 27002:2022, accessed
- Mercator, official project repository, accessed
- Mercator, documentation (introduction, import, API), accessed
- The Open Group, ArchiMate overview, accessed
Frequently asked questions
What is the difference between IT mapping, a CMDB, an asset inventory and enterprise architecture?
An inventory is the list of components and their attributes. A CMDB is a database kept by IT operations that adds the relationships between configuration items. IT system mapping connects the inventory to views that run from business to infrastructure. Enterprise architecture, known in French practice as IT urbanization, is the discipline of planning how the estate evolves: it relies on a map of what exists to define a target and the path toward it.
What is application mapping?
It is the part of IT system mapping that describes applications. In the ANSSI guide it corresponds to the application view: the software components of the information system, the services they provide and the data flows between them. It becomes useful once it is linked to business processes on one side and to servers on the other.
Is IT system mapping mandatory under NIS2?
The word does not appear in Article 21 of Directive (EU) 2022/2555, which lists areas of measures including incident handling, business continuity and asset management. For certain digital entities (DNS, cloud computing, data center and managed service providers, among others), Implementing Regulation (EU) 2024/2690 requires a complete, accurate, up-to-date and consistent inventory of assets. For other entities the detail comes from national law. Either way, managing assets assumes you have listed them.
Which tool should I use to map my IT estate?
The ANSSI guide names no product. It says the choice of more or less specialized tools depends on the target maturity level and on the context, and that dedicated software quickly becomes essential once the volume of data or the number of contributors grows. The tooling has to compile the inventory, produce the views and show the links between them, and control the update process. The guide adds that a simple tool is a major asset.
How many levels or views does an IT map have?
The ANSSI guide proposes three visions (business, application, infrastructure) broken down into six views: ecosystem, business, application, administration, logical infrastructure and physical infrastructure. Separately, it defines three maturity levels for the approach and three levels of detail for objects and attributes. You do not have to build everything: the guide advises working in iterations.
Who is responsible for the IT map?
According to the ANSSI guide, the cybersecurity manager (CISO) coordinates the work when it is centered on cybersecurity, and the CIO can lead it when it is broader. In both cases the objectives are validated by a sponsor who sits on the management board, and business teams are involved as owners of their processes and data. The guide also asks for a specific role in charge of supervising updates.
How often should an IT map be updated?
The ANSSI guide sets no frequency. It asks for regular, structured reviews, names regular update campaigns and an update stage built into change projects as good practices, and says each diagram should carry a date and a version number.