Already in the early phase of a design for ITER, the maintenance aspects should be taken into account, since they might have serious implications. This paper presents the arguments in support of the case for the maintainability of the design, notably if this maintenance is to be performed by advanced remote methods. This structure is compliant to the evolving maintenance strategy of ITER. Initial results of a Failure Mode Effects and Criticality Analysis (FMECA) and a development risk analysis for the ITER upper port plug #3, housing the Charge Exchange Recombination Spectroscopy (CXRS) diagnostic, are employed for the definition of the maintenance strategy. The CXRS upper port plug is essentially an optical system which transfers visible light from the plasma into a fiber bundle. The most critical component in this path is the first mirror (M1) whose reflectivity degrades during operation due to deposition and/or erosion dominated effects. Amongst other measures to mitigate these effects, the strategy is to allow for a replacement of this mirror. Therefore it is mounted on a retractable central tube. The main purpose of this tube is to make frequent replacements possible without hindering operation. The maintenance method in terms of time, geometry and spare part policy has a large impact on cost of the system and time usage in the hot cell. Replacement of the tube under vacuum and magnetic field seems infeasible due to the operational risk involved. The preferred solution is to have a spare tube available which is replaced in parallel with other maintenance operations on the vessel, as to avoid any interference in the hot cell with the shutdown scheduling. This avoids having to refurbish a full port plug and also allows for a more frequent replacement of M1, as we can replace the mirror anytime the vacuum vessel is vented, estimated to be once a year. © 2009 Elsevier B.V. All rights reserved.