A case for static analyzers in the cloud
Résumé
A cloud-based static analyzer runs as service. Clients issue analysis requests through the local network or over the internet. The analysis takes advantage of the large computation resources offered by the cloud: the underlying infrastructure ensures scaling and unlimited storage. Cloud-based analyzers may relax performance-precision trade-offs usually associated with desktop-based analyzers. More cores enable more precise and responsive analyses. More storage enables perfect caching of the analysis results, shareable among different clients, and queryable off-line. To realize these advantages, cloud-based analyzers need to be architected differently than desktop ones. We describe our ongoing effort of moving a desktop analyzer, Clousot, into a cloud-based one, Cloudot.