In addition to the standard offerings, OIT has added NAS-Archive. This offering is an addition to the existing tiers of storage OIT offers. Since 2009, OIT has provided a NAS based CIFS and NFS storage solution. Departments across campus have utilized this storage as an alternative to purchasing and supporting their own storage. This addition of the NAS-Archive solution is targeted at meeting the needs of data that can withstand some reduction in performance and possible uptime with the benefit of a lower cost solution. The primary target data for this type of storage is data from users across campus that is written once and then rarely accessed (aka archive data). By the nature of this type of data, it can withstand some downtime, so the solution providing the data does not need to have as much redundancy (i.e. cost) built in for uptime purposes.
The goal of this solution was to develop a hybrid solution that would still have many of the benefits of a centralized OIT supported solution, but at a cost more in line with what researchers have been paying for their archival storage needs. To be able to meet this price target, the following is a list of differentiators between the standard OIT NAS offering and the NAS-Archive.
|Targeted Storage||Almost any network accessible (CIFS or NFS) storage across Duke University||Any archive type data that will be written once and accessed minimally. If read/write times and/or uptime is a critical factor, the standard NAS should be used instead.|
|Minimal amount required||“None”. Current NAS volumes are as small as 10 Gb.||5 TB initial request with increments of 5 TB thereafter.|
|I/O Requirements||Production NAS storage. 10 Gb connections from the NAS to the core with tools to monitor and identify bottlenecks||1 Gb connection with only minimal monitoring. Data will be maintained on spinning disks and should be accessible at all times, but expectations on time to write and/or retrieve data should be minimal (within reasonable expectations)|
|Monitoring||Multi-level monitoring and alerts that allow for proactive identification of issues and bottlenecks at all levels (network, array, disk, volume, etc.) before they become user impacting issues.||Very minimal monitoring and alerts beyond simple up/down, network connectivity, and/or high level capabilities.|
|Backups||Snapshots for file based recoveries and TSM for disaster recovery with the option of an additional offsite TSM copy sent to Iron Mountain.||No snapshots. TSM is available for disaster recovery at an additional cost.|
|Support||24/7 OIT support with 24/7 “gold level” vendor support and monitoring. Full vendor call home functionality identifies faluts to the vendor’s supportteam often prior to actual failures.||Minimal up/down availability monitoring. 8/5 vendor support. As with all OIT supported services, downtime will be minimizied, however, up to 24 hours of downtime may be scheduled periodically to allow for upgrades and/or rebalancing of the storage arrays.|
|Hardware||High level of hardware redundancy built in (controllers, backplanes, power, etc.) to minimize downtime.||Minimal hardware redundancy built in (power & RAID disks).|
|Access Method||NFS and/or CIFS||NFS and/or CIFS|
|Cost (Actual and Charged)||$.15/Gb/Yr + backups||$.08/Gb/Yr + backups|
Last updated November 1, 2017