Data Storage

Our facility aims to store data in a secure and accessible way. Our storage infrastructure comprises two primary storage areas hosted on Koch Institute servers managed by KI Technology Services (ki-help@mit.edu) and the Bio Micro Center. 

For more details about our data storage practices, please review the rest of this page. You can also find more information about the types of data we work with in our page about Data Types.

Storage Spaces: Shared vs. Archived 

Diagram: Shared storage (teal) is for collaborative data exchange, whereas Archived storage (purple) is for staff-only access. Arrows show the pathways our users and staff have to access data in each space: Users can directly read and write (solid teal arrow) in Shared storage, while staff can read and write in both (solid teal and purple arrows). If users would like to donate, embargo, or remove data to/from the Archived storage, they can communicate directly with staff (dashed purple arrow). Both spaces requre a SMB/CIFS connection from a computer within the MIT on-campus network or remote via VPN. 

Shared Storage 

A Kerberos-authenticated storage space for temporary data storage and exchange. Access is granted to members of the ‘ki-atwai-nfs’ Moira list. Both our users and staff can use this space for a wide range of activities that require storage. Files stored in this space are ideally removed by the user responsible for the data in a reasonable timeframe corresponding with the work they are doing in the lab. Typically this is within one month of completing the work. We evaluate the need to remove data on a case-by-case basis and if we find folders containing files without modification in the last 90 days, we will remove it as needed. Users who need to store data on the server should utilize a folder with your full name or MIT Kerberos ID. This way we can contact you if we find old inactive datasets and need to schedule them for removal.

Archived Storage  

Longer-term storage and staff-only access keep this space secure. Our data archive is designed for secure preservation of datasets. These may be generated directly by our staff or collected with user consent. Data stored here is for internal protocol development only. Uses may include metanalysis, record of failed approaches, and/or training machine learning algorithms for specific use in our instruments – more information at Learning2See. Its access restrictions ensure exclusivity and security for longer-term data storage. It cannot be accessed by other users, only staff. More details about data donations in the next section 

Data Donations: Donate, Embargo, Remove 

We are committed to securely storing any data donated by our users. We manually complete any actions required to ensure attention to the specifics of each dataset. Users can choose to donate data to our facility. Once data is donated, it can be removed or embargoed at the request of the user. More details about each below.

Donate 

Researchers can donate datasets to our facility in coordination with our staff. Donated data will be annotated in a way that allows us to ensure it is managed respectfully and according to any requirements of the user. Donated data will only be used for internal operations that will improve the services we can deliver to our users.  

Embargo 

If you have donated data but need to temporarily limit access, you can request an embargo on that dataset. Upon receipt and confirmation of your request, we will make the dataset inaccessible even for internal use. Please contact us directly so we can establish details of the embargo, including duration and access restrictions. 

Remove 

To remove previously donated data, please contact our staff directly. We will remove your data as needed. Our goal is to securely archive data in ways that avoid removals, but we acknowledge our users’ evolving needs may require this.   

Our Commitment 

We understand the intrinsic value of research data. User data will never be used in a way that detracts its value to the researcher. This commitment is central to our data management practices. All actions—from donation to removal—are conducted with respect for original scientific context and the researcher’s intentions with the dataset.

Please contact us at any time with any questions or concerns so we can address them in a timely way.

Getting Started

To get started, get in touch with us (aiptcore@mit.edu) to apply for access. Once granted, you can follow these instructions to connect from various OS environments. Once connected, please create a folder with your full name or MIT Kerberos ID. This will make it easier for us to contact you if we find old inactive datasets and schedule them for removal.