Definition of ITIL
ITIL is a framework of good practice dedicated to IT service management, a discipline known by the acronym ITSM (IT Service Management). A one-sentence definition of ITIL: it is a structured set of recommendations for designing, delivering, operating and improving the services an IT department provides to its users and customers.
The framework was born in the British government in the 1980s, under the name Information Technology Infrastructure Library. It is neither a standard nor a method to be applied to the letter. It offers a vocabulary and practices from which each organisation keeps what is useful to it.
What the ITIL framework is for
The central contribution of ITIL is a change of viewpoint. An IT department no longer describes itself by its technical components (servers, networks, applications) but by the services it provides: an available email system, an operational workstation, an accessible business application. Each service has beneficiaries, an agreed quality level and an owner.
From this viewpoint flow concrete uses: speaking the same language between support, operations, development and the business, clarifying who does what when a service degrades, governing changes made to production, and measuring the quality delivered rather than the activity of the teams.
The key notions of ITIL
- Service. A means of delivering value to a user without the user having to bear the technical costs and risks themselves.
- Service level. The quality commitment, formalised in a service level agreement (SLA): availability, restoration times, support hours.
- Incident. Any unplanned interruption or degradation of a service. The priority is to restore the service.
- Problem. The cause of one or more incidents. The priority is to understand and remove it.
- Service request. A routine, expected request: an access, a piece of equipment, some information.
- Change. Any modification likely to have an effect on a service. It is assessed, authorised and planned according to its level of risk.
- Configuration. The inventory of the components needed for the services and their relationships, kept in a configuration management database (CMDB).
- Service desk. The single point of contact between users and the IT department.
- Continual improvement. The regular review of services and practices to make them progress.
From lifecycle to value system
ITIL has gone through several editions, and their structure has changed. A long-dominant edition organised practices according to the lifecycle of a service: strategy, design, transition, operation, continual improvement. The edition that succeeded it, ITIL 4, reasons in terms of a value system: it emphasises value co-creation with the business, guiding principles and a catalogue of practices, and it connects more explicitly with agile and DevOps approaches. A later version, ITIL (Version 5), is being rolled out progressively by the company PeopleCert, while ITIL 4 remains available.
The notions presented above are stable from one edition to the next. This is why an organisation can rely on ITIL without following every revision.
ITIL and project management: not to be confused
| Project | Service | |
|---|---|---|
| Duration | Temporary, with a planned end | Permanent, as long as the service exists |
| Purpose | Deliver a change | Provide value continuously |
| Measure of success | Scope, time, cost, benefits | Availability, handling times, satisfaction |
| Frameworks | PRINCE2, PMBOK Guide | ITIL, ISO/IEC 20000 |
The two worlds meet at go-live. A project that delivers an application without operating documentation, without trained support and without a defined service level has kept to its schedule and missed its objective. The ITIL practices of change management and release management serve to prepare this handover. Running the project falls under a method such as PRINCE2.
Nor is ITIL opposed to reliability engineering practices. It says what needs to be managed (incidents, changes, service levels), while SRE offers a way of tooling and automating it.
A textbook case: from incident to problem
In a company, the order entry application freezes several times a month. Each time, the service desk logs an incident, operations restart a component and the service resumes. The incidents are closed within the planned times, and the support indicators are good. Users, however, keep suffering the same outages.
Opening a problem changes the question asked: no longer “how do we restore?” but “why does this keep happening?”. The analysis links the incidents to storage saturation during a night-time batch job. A change is planned to fix the job, tested and then released to production. The incidents stop. The three notions (incident, problem, change) have each played their part.
For implementing these practices in modern operations, see our page dedicated to SRE and observability.