Position from the ST group and the CERN development perspective:

CERN, through the IT department, has been one of the main driving forces behind the Disk Pool Manager (DPM) design, development, evolution and support in the last 15 years. This made the DPM software a pillar for storage solution in a distributed scientific computing research environment in the LHC Worldwide Computing Grid as well as in other sciences.

Given the large scale and the increasing requirements for storage that the LHC experiments need in the coming years, CERN has developed the next generation of open source storage software (EOS) that is currently in Tier-0 production for all the LHC experiments and that is believed to sustain the increased performance demanded for the next LHC run.

CERN is very proud of the existence of a large community worldwide using DPM since several decades. As the community has expressed concerns on the future support that can be expected from CERN on DPM, this documents should summarize the position of the Storage group of the IT department.

The IT department will continue to support DPM on the existing deployments with the present set of features. CERN will continue to ensure to have a minimum set of skills to be able to ensure bug fixes on the current releases and the necessary coordination of the external contributions.

The Storage group is also committed to continue the implementation of the new authentication technology that will become necessary in the next generation of the WLCG software. This effort will be consolidated and released in an upcoming new version.

Finally, there is no plan from CERN to add additional features to DPM, and we encourage the community to continue to invest resources to develop further this package, for implementing additional features and to port it to newer operating systems as necessary.

 

#####

From the WLCG perspective: 

* We should assume that operations with the existing DPM product and support effort at CERN and in the DPM community is sustainable for the duration of RUN-3
* We should initiate a discussion in the next couple of months (Lund was an opportunity, need now to think an alternative) to evaluate storage technologies with a production timescale by HL-LHC. The discussion should consider current storage- and caching-products. 
* The evaluation should consider target functionalities for WLCG (e.g. the DOMA outcomes) as well as long term support and sustainability. 
 

#####