Ten Things IT Should Be Doing to Manage Unstructured Data – But Isn’t
‘To do’ list reduces the risk of unstructured data loss
Dec. 14, 2010 08:45 AM
When it comes to protecting unstructured data, such as spreadsheets, documents, images and other data on file servers, most organizations acknowledge that their existing processes and risk profiles are less than ideal. Unfortunately, IT personnel - rather than data owners - are typically the ones making many of the decisions about permissions, acceptable use, and acceptable access review. And because IT personnel aren't equipped with adequate business context around the growing volumes of unstructured data, they can only make a best effort guess as to how to manage and protect each data set.
Until organizations shift the decision making responsibility to business data owners, IT carries the burden of enforcing rules for who can access what on shared file systems, and for keeping those structures current through data growth and user role changes. IT needs to determine who can access unstructured data, who should and is accessing it, and what is likely to be sensitive.
To help streamline this process, here are 10 must-do actions for IT teams to carry out as part of a daily data management routine to maximize unstructured data protection:
1. Identify data owners
IT should keep a current list of data business owners and the folders and SharePoint sites that are their responsibility. By having this list the ready, IT can expedite a number of the previously identified tasks, including verifying permissions revocation and review, and identifying data for archival. The net effect is a marked increase in the accuracy of data entitlement permissions and, therefore, data protection.
2. Remove global group access control lists (ACLs) like ‘Everyone'
It is not uncommon for folders on file shares to have access control permissions allowing ‘everyone,' or all ‘domain users' (nearly everyone) to access the data contained. This creates a significant security risk, for any data placed in that folder will inherit those exposed permissions, and those who place data in these wide-open folders may not be aware of the lax access settings. Global access to folders should be removed and replaced with rules that give access to explicit groups that need it.
3. Perform data entitlement (ACL) reviews
Every file and folder on a Windows or Unix file system has access controls assigned to it that determine which users can access the data and how, i.e., read, write, execute, and list. These controls need to be reviewed on a regular basis and the settings documented so that they can be verified as accurate by data business owners and security policy auditors.
4. Revoke unused and unwarranted permissions
Users with access to data that is not material to their jobs constitute a security risk for organizations. Most users only need access to a small fraction of the data that resides on file servers. It is important to review and then remove or revoke permissions that are unused.
5. Audit permissions changes
Access Control Lists are the fundamental preventive control mechanism that's in place to protect data from loss, tampering, and exposure. IT requires the ability to capture and report on access control changes to data, especially for highly sensitive folders. If access is incorrectly assigned or changed to a more permissive state without a good business reason, IT and the data business owner must be alerted quickly and be able to remediate the situation.
6. Audit group membership changes
Directory Groups are the primary entities on access control lists (Active Directory, LDAP, NIS, etc.) with membership granting access to unstructured data as well as many applications, VPN gateways, etc. Users are added to existing and newly created groups on a daily basis. Without an audit trail of who is being added and removed from these groups, enforcing access control processes is impossible. Ideally group membership should be authorized and reviewed by the owner of the data or resource to which the group provides access.
7. Audit data access
Effective management of any data set is impossible without an access record. Unless you can reliably observe data use you cannot observe its misuse, abuse, or non-use. Even if IT could ask its organization's users if they used each data set, the end users would not be able to answer accurately - the scope of a typical user's access activity is far beyond what humans can recall. Without a record of data usage, you cannot determine the proper organizational owner for a data set, and neither the unfound owner nor IT can make informed decisions about protecting it, archiving it, or deleting it.
8. Prioritize data
While all data should be protected, some data needs to be protected much more urgently than others. Using data owners, data access patterns, and data classification technology, data that is considered sensitive, confidential, or internal should be tagged accordingly, and protected and reviewed frequently.
9. Align security groups to data
Whenever someone is placed in a group, they get file system access to all folders that list the group on its ACL. Unfortunately, organizations have completely lost track of what data folders contain which Active Directory, LDAP, SharePoint or NIS groups. This uncertainty undermines any access control review project, and any role-based access control (RBAC) initiative. In role-based access control methodology, each role has a list of associated groups into which the user is placed when they are assigned that role. It is impossible to align the role with the right data if the organization cannot verify what data a group provides access to.
10. Lock down, delete, or archive stale, unused data
Not all of the data contained on shared file servers and network attached storage devices is in active use. By archiving stale or unused data to offline storage, or deleting it, IT makes the job of managing the remainder simpler and easier, while freeing up expensive resources.
The principal of least privilege is a well-accepted guideline for managing access controls - only those who have an organizational need to access information should be able to do so. However, for most organizations, a least-privilege model is not feasible, because data is generated far too quickly and personnel change rapidly. Even in small organizations the growing data set and pace of organizational changes exceed the IT department's ability to keep up with access control lists and group memberships. By automating and conducting the 10 management tasks outlined above frequently, organizations will gain the visibility and auditing required that determines who can access the unstructured data, who is accessing it and who should have access. This detailed data access behavior will benefit organizations in a plethora of ways, most significantly securing their data, ensuring compliance demands are met, and freeing up expensive storage resources.