Uncia
FR / EN Demo

Glossary

Application mapping

Application mapping is the practice of recording an organization's applications, who answers for them, and what connects them: data flows, databases, servers. It shows what depends on what before a change, an incident or an audit.

Updated

What is application mapping?

Application mapping is the practice of recording an organization's applications, what they do, who answers for them, and what connects them: data flows, databases, servers. When the work focuses on the links, people call it application dependency mapping. When it focuses on the list and on what to keep or retire, they call it application portfolio mapping.

The French cybersecurity agency ANSSI, in its guide to mapping an information system, describes an application view that covers the software components, the services they deliver and the data flows between them. NIST's Cybersecurity Framework 2.0 asks for two closely related things under asset management: inventories of software, services and systems (ID.AM-02), and representations of authorized network communication and of internal and external network data flows (ID.AM-03).

A useful application map can start as one row and one arrow. The row: the billing application, owned by finance, running on two virtual machines in the production cluster, backed by its own database. The arrow: it receives orders from the ERP every night by file transfer and sends invoices to the e-invoicing platform over HTTPS. That is already enough to know whom to call and what stops if the database goes down.

What should you record for each application?

ANSSI's guide proposes a list of attributes for an application: an identifier and a description, the entities that use it, the entity that operates it, the security lead, the technology type, the application type (in-house development, packaged software, script), its flows, its security needs, its external exposure, the business processes that rely on it, and the databases and logical servers underneath it. Each flow has a sender, a receiver and an encryption flag.

For data, the annex attaches to each database the list of information it holds, and to each piece of information its sensitivity (personal data, for example).

Operations teams almost always add four more fields: a business owner, meaning a named person and not a department; criticality, on the scale the organization already uses for continuity planning; the version in production and its vendor; and the end-of-support date of that version, with the source of the date. The ANSSI guide records operating system versions at the server level; the version of the application itself is a practitioner's field, and you will want it the day you plan a migration.

The guide ranks its attributes in three levels of granularity, from essential to useful. Twenty critical applications described at the first level, with their flows, are worth more than two hundred half-empty records.

How does application mapping differ from mapping the whole IT estate, and from application portfolio management?

A map of the whole IT estate covers every layer. In ANSSI's breakdown it runs from the ecosystem and the business processes down to logical and physical infrastructure. Application mapping is the middle layer, and the one that ties the others together: upward, each application is attached to the processes that use it; downward, to the servers that carry it. An application map without those two attachments is of little use when talking to the business or preparing a maintenance window.

Application portfolio management (APM) looks at the same applications and asks a different question: which ones to keep, merge, replace or retire, and at what cost. It is a decision process, usually run by enterprise architects with finance.

The map describes what exists and what depends on what; the portfolio is where decisions get made. A portfolio built on a declared list, with no verified dependencies, underestimates the cost of every retirement, because an application nobody seems to use may still feed others.

How do you build an application map?

ANSSI's method starts from what already exists: gather the inventories and diagrams you have, then fill the gaps with targeted interviews, collection tools, extracts from applications and internal documents such as continuity plans.

Five sources complement each other. Interviews with application owners give you what no tool knows: ownership, criticality, actual use. The CMDB gives a list and some relationships, often dated. Discovery and endpoint management tools report what runs on the machines. Hypervisor exports (an RVTools file, for example) and AWS or Azure inventories give the machines, their operating systems and their networks, without saying which application they serve. Flow matrices and firewall rules give the permitted exchanges.

The real work is matching those sources on a stable key. An application name spelled three ways in three files becomes three applications. A short application code, repeated in machine names or in a cloud tag, settles most matches. When two sources disagree, keep both values with their dates: the gap shows where the documentation fell behind.

How does application dependency mapping work?

An application dependency can be read at three levels. Between applications: the flow, with its direction, protocol and frequency. Between an application and its components: database, middleware, directory, file service. Between the application and its platform: virtual machines, hosts, storage, network. Impact analysis walks the same levels the other way: this host goes down, which applications fail.

There are two methods. The declarative one starts from architecture documents and flow matrices and tells you what should exist. The observational one starts from firewall rules, logs or agents installed on servers and tells you what actually passes. Neither is enough alone. A month-end closing flow escapes a week of observation, and a flow opened in a hurry on a Friday night is in no document.

The same mistakes come back from one project to the next: leaving out shared dependencies (directory, DNS, backup, job scheduler), even though their failure affects everyone; drawing a flow without its direction; describing applications with no owner; aiming for completeness in the first version and then never publishing it.

Which application mapping tools exist?

A spreadsheet and a drawing tool such as Visio or draw.io are enough to start with a few dozen applications. Everyone knows how to use them. Their limit is well known: the drawing and the list are not linked, so every change has to be made twice.

A CMDB holds the link between applications and machines well when a change management process feeds it. Without that process, it ages.

Enterprise architecture and modeling tools manage a repository, views that stay consistent with each other, and a notation such as ArchiMate. ANSSI's guide considers dedicated mapping software necessary once the volume of data or the number of contributors grows, and notes that few such tools collect the information themselves. Open source tools exist in this family.

Application dependency mapping tools watch traffic or query servers. They see what runs, not what it is for or who owns it.

Source-derived mapping rebuilds the map from existing exports and inventories. It is as good as those sources, and it has to say what it is missing. Choosing an application mapping tool therefore starts with one question: who will keep the information current, and from what.

What Uncia adds, and what it does not do

Uncia belongs to the last family. It reads your sources read-only, 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 changes nothing. From that base it generates Architecture, Infrastructure and Network maps, from the business domain down to the firewall rule, a record for each element with its sources, and the impact of a shutdown. Maps export to ArchiMate, draw.io, PNG or SVG. The knowledge base updates with every collection.

Uncia is not a monitoring tool: it works from what your sources contain and tells you what it is missing. What no source says, such as the owner of an application absent from the CMDB, still has to be filled in by a person; when the language model proposes a value, someone accepts or rejects it. Uncia reads your CMDB and your architecture models and shows where they diverge from the other sources, without standing in for them. More on the page about maps generated from your sources.

Sources

Frequently asked questions

What is the difference between application mapping and application dependency mapping?

Application mapping describes applications and everything around them: owners, data, hosting, dependencies. Application dependency mapping is the part that deals with the links: which application talks to which, over what protocol, in which direction, and on which infrastructure each one runs.

Which applications should you map first?

ANSSI's mapping guide recommends starting with the most critical or most exposed systems and extending in iterations. For a large estate, it advises limiting the first version to a few views.

Is a CMDB enough to produce an application map?

It is one source, rarely the only one. A CMDB holds what was declared. Hypervisor exports, cloud inventories and flow matrices show what was deployed and permitted. The map comes from reconciling the two.

How often should an application map be updated?

ANSSI's guide sets no frequency. It recommends regular update campaigns involving the people responsible for each scope, and asks that every diagram carry a title, a date, a version number and a legend.

Does any framework ask for application mapping?

NIST's Cybersecurity Framework 2.0 asks for inventories of software, services and systems (ID.AM-02) and for representations of authorized network communication and network data flows (ID.AM-03). Article 21 of the NIS2 Directive lists asset management among its risk-management measures and does not use the word mapping.

Every week, an article to know your IT estate better

Architecture documents, end of support, NIS2, CMDB, migrations: methods you can apply with or without Uncia.

Get a call from our team

Thirty minutes with a member of our team, remote or on site: Uncia on a demonstration instance, then on part of your own estate if you wish. We reply within 48 hours.

Request received

Thank you. A member of our team will reply within 48 hours to set a date.