In my last several posts on members of the IBM Spectrum Storage family, I’ve talked a lot about the growth in data that is forcing IT managers to think differently about storage. Much of the data growth is being fueled by new workloads and their seemingly insatiable need for data to process. But in many enterprises, even more of the data growth is the result of simply keeping data around – stuff like regulatory archives you have to keep and asset archives you just want to keep. As this type of inactive data continues to grow, questions arise about the best way to store and manage it without losing many of the conveniences of online storage. This is where IBM Spectrum Archive excels, balancing the convenience of file system access with the ultra-low cost and flexibility of tape.
My daughter is a student at Texas A&M University, and is her daddy’s pride and joy. She’s part of the millennial generation, a demographic who can’t remember a world without almost unlimited access to digital information. I don’t recall a mental image of her without a smartphone, tablet or laptop in the picture. For her, the answer to every question is found in those devices. Every daily activity – listening to music as she runs, working on school assignments, talking with friends, capturing pictures or videos of important events, watching a movie, and a myriad of other apps – is done with those devices. In her world, a television set is just a bigger screen for Chromecast and a USB flash drive is something you use when Dropbox or Google Docs won’t do the job. Like, for example, when she needs to carry a document into a print center or load a presentation on her professor’s laptop to present to the class. It’s her version of data interchange.
The digitally native world my daughter lives in is what’s responsible for the tremendous data growth we are seeing in IT. These are the customers that business is trying to reach and the employees who are increasingly developing new workloads. They don’t use terms like “data archive”, but they do it for stuff like pictures that they just want to keep – and they use places like Dropbox that don’t really work for most corporate applications. They also don’t say “data interchange”, but they have grown up doing it on media like USB flash drives that don’t have the scale required by most businesses. As they bring their expectations into business, what they need is a super-sized alternative with reliability, security and duty cycle fit for business IT. That’s IBM Spectrum Archive.
In my post IBM Spectrum Scale – Built For Efficiency, Optimized For Speed, I introduced a software defined file and object store that includes rich policy for optimizing data placement across tiers of online storage. With Spectrum Scale and common building block storage, a single namespace can house something like 9 quintillion files. In my daughters mind, that’s like a super-sized business IT alternative to the storage on one of her personal devices. Think about it for a second. For many datacenters, 9 quintillion files could represent all their data – all of it – neatly organized in a single namespace so it’s easy to find and easy to manage. While it’s cool for applications and people to be able to access data through a neatly organized file system interface, not all data needs to be housed in prime online real estate. It’s like the pictures on my daughters’ phone – easy to find, but when space on that prime real estate gets tight, the older pictures need to be “archived” someplace else. The super-sized business IT alternative is IBM Spectrum Archive plugged-in to IBM Spectrum Scale.
Here’s the scenario.
- You’ve got a pile of files in a single IBM Spectrum Scale namespace and you’ve leveraged policy to optimize data placement across online storage like Flash and spinning disk.
- Some of your data needs to be archived (you’re a regulated business or you’ve just got intellectual property records you want to keep). So you plug in IBM Spectrum Archive to add a tape tier. The same policies are now extended to automatically place the right data on tape. In doing the selection, there’s all kind of granularity in the metadata you can write the policies against.
- A tape tier requires some unique handling that Spectrum Archive seamlessly provides. Tapes can be grouped into pools, files can be replicated across multiple pools, and media is reclaimed as files move on.
- The wow factor in all this is that workloads and users who had their files in Spectrum Scale don’t lose the neatly organized file system interface when files are archived to tape with Spectrum Archive. It’s transparent. All the files are still there in the common namespace.
- For data interchange, a tape and its files can be exported from the namespace – basically checking those files out. Conversely, a Spectrum Archive tape can be checked in too (imported), and its files automatically appear in the Spectrum Scale namespace. Encrypting interchange tapes is also an option with the keys managed by IBM Security Key Lifecycle Manager.
Tape entered the IT landscape in the baby boom generation. With software defined storage and Spectrum Archive, it continues as a core storage media for millennials.