To request an RDF-Isilon Network share, please use our web form to Request Help with the Center for Research Computing Resources

For information about general purpose data storage platforms at Rice, please see Rice Storage, File Sharing and Backup Solutions

The Center for Research Computing operates an on-premise 2.9 petabyte (PB) networked data storage appliance known as the Research Data Facility (RDF).  This system is highly resilient, flexible enough to meet a wide range of use cases, and scales easily to meet future data storage needs. The CRC can allocate RDF network shares to faculty researchers, with a 500GB subsidized allocation and cost recovery for additional utilization beyond the 500GB limit.  

RDF Specifications & Technical Information

• • Dell/EMC clustered storage appliance
  • Highly redundant, can tolerate multiple disk failures/head failures without data loss
    (See our note on disaster recovery and backup services below)
  • On-premises, managed by CRC in Rice's Primary Data Center
  • Fast network paths to data center based client systems
  • Real-time, shared access to data from multiple endpoints
  • Please note:  RDF Isilon is suitable for workstations or laptops which need access to an interactive, shared storage area for research data.  It is not intended for high performance workloads, use on multiuser systems or servers, or in support of server infrastructure.

Access to Shares (SMB/CIFS and NFS file sharing protocols)

RDF shares are normally provisioned to support the SMB/CIFS file sharing protocol (i.e., Windows file shares)

• • Authentication using NetID/NetID password and Active Directory credentials
  • OIT retains full administrative rights, to manage permissions on behalf of researchers
  • OIT will grant read/write/modify permissions to researchers’ shares with the approval of the research PI or a designated delegate.
  • To request access, please email helpdesk@rice.edu, and be sure that the approving PI is copied on the request.
  • Complex file permissions and sub-group structures are not supported, however, PIs may request a separate private group for their own use within a share.
  • SMB/CIFS is supported on Mac and PC platforms, with limited support available for Linux distributions
  • CRC strongly recommends against using SMB/CIFS for the following applications:
       - Supporting Unix/Linux network infrastructure (i.e., shared home directories, shared applications)
       - Multi-user Unix/Linux systems

For detailed instructions on how to access RDF shares via SMB/CIFS, please see the RDF User's Guide.

RDF shares can be provisioned as NFS v3 shares on a case-by-case basis when SMB/CIFS is not feasible due to client operating system limitations or application limitations.

• • NFS v3 is unauthenticated, and carries a higher security risk since share permissions can be overridden by a local user with administrative rights
  • NFS v3 shares can be compromised by an NFS client spoofing an IP address which has been granted access to the share
  • Users should be aware that NFS v3 shares cannot be accessed by any systems which are not included in the access control lists
  • Mixed Mode (i.e., SMB/CIFS plus NFS v3 on the same share) is not supported; however a researcher may request a separate SMB/CIFS enabled share. 
  • Shares can be provisioned as NFS v3 once the researcher acknowledges and accepts full responsibility for the security risk
  • Client machines requiring access to NFS v3 shares must be added to access control lists maintained by CRC/OIT
  • To request an NFS v3 share, a Request Tracker ticket should be filed with the help desk to document approval.

RDF Isilon Subsidy Eligibility 

Research storage allocations will be granted to all tenured faculty, tenure-tack faculty and research faculty; these roles are defined by Rice Policy 201.  The first 500GB of storage will be subsidized, and utilization above the subsidized level will be charged back to the researcher.  
Research storage allocations for other research groups not defined above will be handled on a case-by-case basis.

RDF Isilon Charge Back Rates

The current rate for RDF-Isilon is $70/TB/year  (7 cents/GB/year).  The storage amount billed in a month is based on the average storage space used throughout the month. Your storage usage is measured in "GB-Month," which are added up at the end of the month to generate your monthly charges.  For example:

• • A user requests a new allocation in May 2018, and their May average utilization is 290GB.   Since their utilization is below 500GB, there would be no charge for May
  • The user’s average utilization in June 2018 rises to 950GB during a major project.   The monthly charge for June would be:  (950-500)*.07/12 or $2.56
  • The same user finishes their project and removes several large datasets, so their average utilization for July goes down to 600GB.  The monthly charge for June would be:  (600-500)*.07/12 or $.57 

Quotas and Scaling

When a researcher is granted an allocation, a warning (soft) quota will be set at 90% of the hard quota; for example on a 500GB share, the soft quota will be 450GB and the hard quota 500GB.  When the soft quota is reached, the system will send a courtesy email to the researcher warning them that they are getting close to the hard quota.  The hard quota will ensure that the user does not exceed their subsidy and receive unexpected charges.  When the hard limit is reached, an additional email notification will be sent to the researcher, and a help desk ticket will be opened to alert the CRC.  The researcher may then authorize that the quota be removed, with the understanding that they will be billed for any monthly average use above 500GB.  Once the quota is removed, RDF-Isilon storage allocations will scale automatically to meet the demands of users, without additional intervention.  However, to help researchers manage the costs of their storage, they may optionally request the CRC to set soft or hard quotas at a higher limit.

Backup and Disaster Recovery 

Snapshots

• • Each RDF-Isilon share will have nightly snapshots enabled by default
  • Users can locate files in the snapshot directory and restore them within the 24 hour snapshot window
  • Snapshots are *not* a backup or disaster recovery tool, they are intended to allow users to recover accidentally removed files from the last snapshot
  • Data preserved by snapshots are included in utilization averages for cost recovery
  • Snapshot policies can be customized to meet researchers’ individual requirements, through consultation with CRC

Fee-Based Backup/Disaster Recovery Services

• • Backup to Amazon S3 Glacier - Deep Archive
  • Off-site, disk-to-disk disaster recovery
  • Note that Rice University's Amazon S3 Deep Archive vaults are located in the US East (Ohio) region
  • Researcher is responsible for associated AWS Glacier charges.  For current AWS Glacier charges, see: https://aws.amazon.com/glacier/pricing

Security

The RDF is designed to securely accommodate non-regulated data.  If your data requires additional security precautions to ensure regulatory compliance, please contact the CRC to discuss your requirements.

Data Transfer  

Globus Connect/Science DMZ  offers researchers a method of transferring large data sets between participating research institutions.  CRC has integrated the RDF/Isilon appliance as a Globus Connect endpoint, with access to the facilities of the high-speed Science DMZ-Internet2 on-ramp