Combine flash and tape and what do you get? FLAPE
It’s a new emerging storage architecture for long term data retention that combines FLash drives and tAPE = FLAPE. Using flash, you get instant access to data that was most recently accessed (written or read) or to metadata to help you quickly search your vast amounts of data that has been archived. With tape as a back-end storage platform, you have a cost-effective storage solution that blends the performance benefits of flash with the economics of tape.
The combination of tape and flash, for large objects/files, is going to offer
not only lower costs, but also much higher performance than spinning disk-based
alternatives. This changes the dynamics of storage for long-term data retention
and so-called big data lakes. The key to this approach will be surfacing
metadata, currently buried on tape cartridges, to a flash layer that can signal
the location of desired data on the tape. Combined with linear tape file system
technology, we believe this approach will deliver better business value for the
right use cases. This contrarian view is a function of: 1) Lack of investment in
high speed disk technology, specifically head tech; 2) The far better bandwidth
improvements capable with tape relative to spinning disk and 3) The economics of
flash relative to disk.
Using an Archive Manager Software with a tape library. This solution utilizes cache storage sitting in front
of tape and presents a virtual mount point to network clients. These clients
simply mount the volume and access it like any other volume. File writes made to
the volume are first staged in the cache and are written to tape when the cache
fills to a predetermined level. Stubs are left in place at the file location and
the data is written to tape using Linear Tape File System (LTFS). For on-going
protection, the database information in the flash cache is frequently
written to tape.
While some users may choose to use disk as the cache storage, you can simply
substitute your disk system with flash drives and instantly increase your
performance. The flash drives can be installed internally to the host machine or can be in an external array.
With flash drives, lab testing showed a
performance increase of 100% from 135 MB/sec to 270 MB/sec of throughput per
virtual volume in the worst case scenario (simultaneous cache writes, cache
reads, tape writes & database updates). Read-and-write operations to the cache
without the tape overhead is even faster. By adding more volumes to the host, you can easily scale your aggregate throughput to over 2 GB/sec (final
throughput dependent on host specifications).
Depending on the specific software that you use to manage your
archive data, you can also use
flash drives to host the metadata information that your application uses for
data queries. This provides for quick searches to help pinpoint the information
you need to retrieve. If the data is in the flash cache, it will be ready
immediately and can be read quickly. If it’s on tape, Archive Manager will
automatically retrieve it with only a slight delay while the cartridge is loaded
in the drive. Once it’s pulled from tape, it will stay in the flash cache for
subsequent retrievals based on access and will be overwritten only when new data
is accessed and there is not enough room in the flash cache. For example, if a
news story breaks and you need to pull video clips or photographs to supplement
a report, these items will be readily available for other editors for additional
news stories that may use the same content.