Big datasets used to be the exclusive-domain of only a few verticals – Financial Services, Oil and Gas, Life Sciences to name a few. These organisations are constantly creating information and making new discoveries that require cataloguing. However, the explosion of the online world, new legislation around data retention, the move to digital content, as well as the growing realisation that a company’s data may yield insights into all manner of trends amongst its customers, are changing the game for organisations of all sizes and industry sectors.
Mid-sized organisations across all verticals are tasked with storing an increasing amount of data and, equally important, store it for longer. However, while mid-market organisations are confronted with similar challenges to those facing larger enterprises – there are some differences in the resources they have at their disposal. Mid-market storage systems must take into account the fact that IT departments in these organisations are often small and under-resourced.
Finding the right balance between cost, performance, reliability and ease of management is imperative when making decisions about the kind of infrastructure required to support the growth in data. Given its traditional strengths around cost and density, coupled with recent improvements in reliability and automation it’s almost certainly time to think about the role tape can play in meeting long-term data storage needs.
There is significant evidence that supports the argument that for mid-market organisations tape is less staff-intensive to manage than disk, due in part to the leaps that have been made in tape automation technology. With automated tape libraries, backup jobs can be scheduled to run automatically, significantly reducing the IT time required for organisations using a stand‐alone tape drive.
Tape libraries aimed at the mid-market now have simple to use graphical interfaces that offer simple access to all the automation features. Features such as lifecycle automation, key management and media health monitoring, can be easily configured through touch screens or simple remote web browsers. More advanced mid-range libraries also provide the option for Assisted Self-Maintenance – end-users can readily replace components when issues arise under instruction from technical support, rather than waiting hours or days for on-site engineers to arrive.
In regards to long-term capacity requirements, tape presents the perfect balance between cost, capacity and scalability. Unlike Disk, tape libraries can scale in a very modular way and can be added in small quantities, increasing over time by small increments without requiring a significant overhaul or disruption.
Furthermore, a single library can be used to manage many tapes, even beyond the slot count, which can be achieved by taking tapes out of the library when it is full, adding new tapes as needed, and by re-using existing tapes where the data is no longer needed. Capacity on Demand features in mid-range libraries also permit slot increases of as little as five at a time, which is perfect for organisations whose IT budgets require more of a slow and steady approach to expenditure over more erratic but larger-scale investments.
Reliability is also an issue for mid-range organisations and is an area where disk has not fared well over recent years. One analyst from Data Mobility Group suggests that organisations should start to consider disk refreshes after the three year mark. While typical disk life varies based on usage patterns its reliable life compared to tape is short, as modern tape technologies can typically store data for 30 years or more.
Coupled with the fact that tapes can be stored offline without consuming power or valuable data center space, tape is clearly a more viable long-term storage platform than disk. Data protected on disk‐only is further vulnerable to the network environment in a way that tape isn’t, simply because disk is online.
One of the commonly cited concerns about committing data to tape long-term has been ensuring the data can be read if it needs to be accessed. After data has been written to tape, confirming the integrity of data stored onto tape media has only been possible through labour and time-intensive manual testing of tapes.
Today, however, Data Integrity Verification allows organisations to check the health of tapes before they are used, perform a quick scan of data immediately after it is written to tape, and even check an entire tape to ensure it can be read. The IT manager can then set triggers to monitor the health of tapes over time, ensuring data is retrievable.
Aside from the reliability and manageability advantages offered by tape there is a significance financial benefit. In terms of straight up acquisition costs per Giga Byte (GB), tape comes in significantly cheaper than disk at around 75% for comparable capacity. Furthermore, numbers vary based on each individual use-cases, but the Clipper Group puts the Total Cost of Ownership (TCO) of disk over a twelve year period somewhere around 15 times higher than tape.
This is based not only on the higher cost per gigabyte of disk, but also the increased density and lower power and cooling costs associated with tape. While improvements in technology such as de-duplication and thin-provisioning are helping customers to lower the TCO of their disk infrastructures, there is still a significant gap between the short-and long-term cost of disk and tape.
Analysts and industry observers disagree on many things, but where data growth is concerned there seems to be unanimous acceptance that it is a major issue. The reality for most organisations is that data growth over the next five to ten years will be significant but unpredictable.
Organisations know that in all probability they will have to budget for and provision significant amounts of resource, but estimating how much is tough to work out. When selecting the right technology and architecture, management and automation can be decisive factors, but one of the biggest decisions an organisation will have to make is choosing between disk, tape and a mixture of the two.