It’s been a few weeks since we’ve dug around storage technologies on this blog, so it seems a reasonable point to discuss two of the strangest terms bandied about in the storage world over the past year, or so.
So, just what are flape and floud? Although both rejoice in their own Urban Dictionary entries, these are not the flape and floud we’re looking to discuss here. Both are a cunning combination of flash and something else. Although the cost of flash storage has tumbled over the past couple of years, it’s still not super cheap, especially if you’re looking towards larger volumes, hence the search to combine flash with other, cheaper bulk storage options.
With the fall in flash costs, as well as its key USP of no moving parts, means that disk, as a storage medium is now moving towards an awkward position. The total cost of ownership of disk looks to be increasingly challenging to justify, especially in situations where assets are potentially going to be sweated out much longer than the conventional three year initial warranty period, maybe to five, or seven years. There is no trumpeting the benefits of ‘flisc’ in the future roadmap of storage!
The very slow rate of improvement in performance of disk-based systems has been coupled with a steady climb in the capacity and speed of access of tape, with LTO-7 being announced at the end of 2015. The rise in capacity per cartridge now means, with LTO-7 each unit delivers 16TB of capacity – ten times more than an LTO-4. In addition to this step up in capacity, the new standard has added more heads to read and write to tape simultaneously. Despite the repeated announcements of the death of tape, it seems to continue to go from strength to strength, to the point that we are looking at a brave new world of tape as data storage, becoming extremely effective when combined with flash storage, which is used as the indexing database to create…. Flape!!!! (pause for breath)
Flape might be particularly appropriate in situations where large media files, such as video etc. are stored. It’s an ideal solution for a media library, where files might not need to be accessed for a long period, but it’s important systems can navigate to them swiftly, when they’re needed. It also has a great use-case where large volumes of academic research data need to be retained, but might not be referred to very often. With the increase in the number of heads available to read from tape, combined with the high speed identification of the metadata, once files are located, they can be streamed off pretty much as quickly as a file could be read from disk, it’s just the start of the read that might take slightly longer.
It is very unusual for any discussion about storage technologies to get this far through a post without mentioning the C-word, so let’s remedy that immediately: Floud is a glorious combination of flash for indexing and metadata, combined with cloud as the extensive data storage medium. This is an interesting proposition for people reviewing their backup solutions, especially if there is an Iron Mountain tape off-siting solution in place. Whilst tape off-siting is a useful way of securing data, it does rely on the physical movement of media to provide that life-saving data recovery solution that’s likely to be needed if a disaster recovery situation is ever invoked. Moving to a – ahem – floud solution means your recovery data is held, ready to be retrieved from the cloud, without the logistics and timescale associated with a recovery from cold, off-sited data, especially if this hybrid solution was combined with a DR as a Service (DRaaS) solution, such as the Silverstring Predatar DRaaS.
The cloud storage container pools introduced in 7.1.3, along with the inline compression introduced in 7.1.5 means that floud solutions to provide your off-site backup images is now very much an option worthy of consideration. If you would like to discuss this further with me, then please fill in the form below.
