Cincinnati Children's is implementing a new Research Data Storage (RDS) system that replaces the current "Celerra" servers (G: and H: drives) and other internal file servers such as "Searchers."  This new system also will accommodate data that are currently stored on external hard drives. Researchers are strongly encouraged to take advantage of the RDS, which provides centralized and backed-up storage at a low or no cost to the investigator. The Division of Biomedical Informatics is spearheading the rollout effort; the Cincinnati Children's Department of Information Services has operational control.

If you have general questions about the RDS that are not answered on this page, please feel free to email Dr. Michal Kouril, Director, Research IT Services, or Dr. Michael Wagner, Faculty Liaison, Biomedical Informatics.

RDS goals

• To make centrally managed, secure data storage easily and transparently available to all researchers at Cincinnati Children's
• To manage growth of the storage system so that it is predictable for the institution
  
• To provide some basic amount of storage for free and make larger allocations available for a very competitive and attractive charge-back.
  

Institutionally provided quotas (free of charge to investigators/divisions [1])

• 10GB per user (the "H: drive"). This replaces the current H: drive on the Celerra system with one that is limited by a 10GB quota. Investigators may request quota increases (see terms below). The H: drive is only accessible to the investigator him- or herself and is meant for personal data such as email etc.
  
• 100GB shared partition per division (the "G: drive"). This replaces the current G: drive on the Celerra system, and it is meant to store ALL data related to divisional business and administration. The use of the remainder of the quota is left at the discretion of the division and can be increased for an annual charge (see below).

Project directories for collaborations

• Investigators who wish to allocate data storage for their research projects will be able to send a request and get a project drive assigned to them within a few days.
  
• All storage for project directories will be charged back (see below for rates).
• Investigators will be able to control who can access data on their project drives. Permissions can be given to any collaborator within institution (not limited to the investigator's division).
• BMI is currently developing an online tool for requesting increases to personal and division-shared quotas, and for requesting new project shared storage. In the meantime, please submit requests with details on how much space you need and who needs Read-only or Read-Write access to the space to the BMI Help Desk.

Current charge back structure [1]

• $1 / 1GB / year -- backed-up storage (read-write storage with two months worth of daily backups; for 12 months worth of daily backups, add $1/1GB/year)
• 50c / 1GB / year -- storage without backup (read-write storage without daily backups)
• 50c / 1GB / year -- archive (one-time read-only snapshot)

These charges do not cover the true costs of this storage to the institution but are rather a management instrument that should entice divisions and investigators to use data storage responsibly. We want to encourage investigators to periodically clean up their data and at the same time minimize unnecessary duplication.

Support

All components of the system are supported 24x7 by the Department of Information Services (IS). Storage quota change requests are fulfilled within eight business hours (Monday through Friday, 7 am - 7 pm). New storage allocation requests are fulfilled within two business days.

What RDS replaces

RDS replaces the current Celerra (\\celerra1 and \\celerra2) servers (G: and H: drives), as well as other internal file servers such as Searchers and directly attached hard drives such as USB and Firewire hard drives.

Migration plan

Information Services will automatically migrate user directories with less than 10GB of space. BMI will assist users with more than 10GB to find a proper placement for the data. BMI also will assist divisions during the migration of the shared partition. Users and divisions will be notified in advance of the migration. For a tentative schedule, see our RDS transition schedule page.

Archiving options

In addition to the charge-back option of a one-time, read-only snapshot, BMI can work with you to evaluate options such as a DVD or external hard drive achival copy. Contact the BMI Help Desk if you are interested.

Performance

Performance varies based on the number of connected users and their activities. Maximum read performance is 40MB/s , while maximum write performance is 20MB/s [2].

Supported access protocols

RDS currently supports Windows file sharing access from Windows, Mac and Linux. NFS support is planned for the second half of 2008.

Underlying technology

• Vendor: EMC
• Model: NS80 with Centera
  
• Backup is based on daily snapshots
• Highly redundant storage (two sets of everything for the disaster recovery location)
• Consists of many RAID5 (single parity disk) and RAID6 arrays (dual parity -- can withstand two hard drive failures in the array)
• Raw HDD space: about 200TB
• Usable space:  about 60TB
• Quotas management software by Northern Parklife (web accessible quota information) is under evaluation
  
• Permissions management software by Varonis DataPrivilege (web accessible permissions requests and grants) is under evaluation
• Auditing software by Varonis DatAdvantage is under consideration

[1] Subject to periodic review

[2] Measured using Windows file sharing client on Windows, Mac and Linux