Skip to main content
Fig. 12 | BMC Bioinformatics

Fig. 12

From: Watchdog – a workflow management system for the distributed analysis of large-scale experimental data

Fig. 12

Comparison of Watchdog with other WMSs. Comparison was performed using features grouped into the categories setup, workflow design, workflow execution and integration. Workflow is abbreviated as workf. in this table. Integration refers to the integration of new data analysis tools into the particular WMS. Footnotes: 1six non-free extensions are available; 2since version 2.4.8, rules can also explicitly refer to the output of other rules; 3explanation: includes a way to automatically run a predefined workflow for a variable number of replicates based on filename patterns; 4have to be created manually in the web-interface from uploaded files; 5explanation: finished steps of the workflow can return variables that are used by subsequent steps as input; 6can only return the names of output files; 7other supported executors: Watchdog: new executors can be added with the plugin system, Galaxy: PBS/Torque, Open Grid Engine, Univa Grid Engine, Platform LSF, HTCondor, Slurm, Galaxy Pulsar, Snakemake: can also use cluster engines with access to a common file system and a submit command that accepts shell scripts as first argument, Nextflow: SGE, LSF, Slurm, PBS/Torque, NQSII, HTCondor, Ignite; 8non-free extensions for SGE or dedicated server support are available; 9custom executors for cloud computing services can be created using the plugin system; 10Watchdog: HTTP/S, FTP/S and SFTP by default, can be extended to any remote file system with an implementation of the FileProvider interface from the Commons Virtual File System project, Galaxy: Object Store plugins for S3, Azure, iRODS, Snakemake: S3, GS, SFTP, HTTP, FTP, Dropbox, XRootD, NCBI, WebDAV, GFAL, GridFTP. Nextflow: HTTP/S, FTP, S3; 11a hard-coded error checker triggered on keywords ‘exception’ and ‘error’ in standard output and error is provided; 12depends on the node implementation and left to developer; 13explanation: usage of local storage during distributed execution in order to avoid unnecessary load on the shared storage system; 14direct integration of python code is possible; 15own scripting language available; 16explanation: describes the concept used to separate workflow definition and functionality (e.g. Watchdog’s modules) in order to allow easy re-use of functionality; 17modules can include binaries in the module directory or automatically deploy required software using Conda, Singularity, Docker or similar tools available on the used system

Back to article page