Virtual Tape Advanced Recall (VTAR)

Slash run times for batch workloads

Virtual tape cache is a finite amount of storage and virtual tape volumes are eventually migrated to physical tape for different reasons. When jobs request virtual tape data residing on physical tape, a recall must be performed and the volume containing the requested data set must be mounted and the data set migrated back to disk cache. This recall process causes a delay for all jobs requesting a data set which is not in cache. Virtual tape recalls utilize time and resources regardless of which virtual tape device you use. VTAR can pre-stage your migrated virtual tape data and eliminate virtual tape recall wait time.

Think you don't waste a lot of time on virtual tape recalls?

Though varying from one manufacturer to the next, an average virtual recall mount time is about 200-300 seconds per recall. Even if the data center's cache hit rate is just 94% daily, the time savings can be significant. If 2,000 virtual mounts are created daily and six percent of those mounts (120) are recalls, at 300 seconds per recall those 120 recalls would cause 10 hours of wait time.

Drastically reduce periodic cycle times

Workloads that are not run every day will require the most volume recalls since they possess the greater chance of having required tape data sets migrated from cache due to inactivity. The windows for all your weekly, monthly, etc. workloads will suffer due to the additional processing time necessary to service the recalls. VTAR eliminates this wait time and allows these periodic cycles to run faster.

Postpone or eliminate cache upgrades

By optimizing the recall of virtual tape datasets, VTAR helps you make more effective use of your existing virtual tape cache and can eliminate or postpone expensive virtual tape hardware cache upgrades.

No additional batch processing overhead

Before recalling a virtual tape dataset, VTAR will first verify that the dataset hasn't already been recalled to cache by another process or previous job. There is no unnecessary overhead.