Two topics – really on opposite ends of the storage spectrum – captured my attention this week. The first had to do with an announcement that Imation made this past Wednesday regarding it being the first and only company currently licensed to manufacture LTO-5 tape media. The other had to do with cloud storage and some of the conversations that I continued to have with various providers in terms of how ready (or not ready) cloud storage is for prime time.
First, I briefly spoke yesterday to Will Qualls, Imation’s Product Director for Tape Products, about its recent announcement as well as what Imation is experiencing in terms of momentum around tape in general and LTO technology specifically. The significance of the announcement is that so far Imation is the only storage vendor that is currently licensed to produce the LTO-5 media in anticipation of LTO-5’s early 2010 launch. Since the specifications on how to manufacture LTO-5 tape media are not made public and only released to manufacturers “approved” by the LTO consortium, this would seem to give Imation a significant head start on its competition. However Qualls declined to comment as to exactly why Imation is the only manufacturer to receive this approval status.
Qualls and I also discussed the tape market in general and he made a couple of interesting comments in regards to tape’s growth. When asked about how deduplication was affecting the tape market overall and LTO specifically, he said the tape industry has definitely seen a reduction in the number of units shipped but whether that is due to increases in tape capacity or the introduction of deduplicating disk systems into backup processes, he could not really say for sure. However Imation is definitely seeing tape increasingly used in an archival role through Imation anticipates that deduplicating storage systems will show up there sooner rather than later.
While LTO has become the predominant tape format over the last few years, LTO is still looking to broaden its reach. Qualls mentioned the IBM 3590 tape format as a specific area where it has seen some success recently in displacing that format with LTO. I then asked him if LTO had any plans to support mainframe connectivity and on this point he hedged a bit. Since IBM is part of the LTO consortium and also controls the mainframe market, he currently is not aware of any plans to attach LTO to the mainframe.
I guess this makes sense from IBM’s perspective. After all, why would IBM want to undercut itself my giving its mainframe users a more economical option for tape backup? Let them use 3590 media and pay more. Sometimes I think it is a miracle the mainframe has survived as long as it has with that mentality.
In any case, back in the open systems side of the world where it is a rarity for anyone to survive, I had number of engaging conversations with EMC, NetApp and Mezeo Software, a relative newcomer to the storage scene, on the topic of cloud storage. Again, these interviews were done in conjunction with an article I am writing for SearchStorage.com so I am only sharing a couple of tidbits of information that I gathered that are not germane to the article:
- It is important for anyone using contemplating the use of cloud storage to understand that it uses object-based storage. Thia means to access and retrieve data stored to the cloud, applications have to use web access protocols like SOAP and REST. This may preclude some applications from using cloud storage as a storage option in the near term until they add support for these protocols. It also means some storage products will need software overlays in order to support these APIs. I talked about this last week in a blog where Mezeo Software and Permabit were cooperating in such a fashion. In that vein, this past Wednesday they formalized their relationship with an announcement that they were entering into a partnership. Sources in the industry also tell me that Mezeo is having similar conversations with other competing archiving providers.
- The upside of storing all data as an object is that it creates new possibilities for storing and accessing data. Mezeo tells me that this can eliminate current file system constraints such as using hierarchical file systems and allows for the creation of large fields in which to store metadata. This will eventually allow applications to tag objects in many and multiple ways and potentially greatly improve search capabilities of data archived to the cloud. Search engines can first search the metadata and never have to touch the actual archived data unless it actually needs to be retrieved to satisfy a retrieval request. Since many cloud service providers charge for bandwidth and access charges, this stands to significantly lower the ongoing operational costs associated with storing data in the cloud.
Finally, there are a couple of pieces of industry news that I picked up over last few weeks but that I can not yet publicly share. First, I was recently informed that the CEO of a company in this industry has stepped down via an email from the CEO. The company for which he used to work verbally confirmed his departure when I called them. I promised I would wait to blog about it until there was a press release on its website but, as of last night, there was no announcement.
The other tidbit of news I picked up is that a storage hardware company that has been struggling for some time and looking to sell its technology may have found a buyer. Again, this is not definitive but I may be able to share more next week as to who the company and its buyer are.