Cloud Storage Services Evaluation
This document compares several cloud storage solutions freely available to faculty, staff, and students at Johns Hopkins University. Table 1 compares the services from the perspective of collaboration, sharing, and/or syncing of data for research. This is reflected in the comparison criteria. In addition, some best practices for working with research data in an online collaborative environment are outlined.
This comparison is for informational purposes only and does not constitute an endorsement by Johns Hopkins Data Management Services.
Comparison Table – Basic Information, Syncing, Collaboration, Other Capabilities
Although there are a wide variety of cloud storage services available, four options based on popular services that are available without charge to the JHU community are compared in Table 1: JH-Box (based on Box), JHU OneDrive (based on Microsoft OneDrive), Dropbox, and Google Drive. The focus is on comparing attributes that affect the usability of the service for the purpose of temporarily storing and syncing personal or research files. Additionally, attributes related to conducting collaborative research are evaluated. Although Dropbox and Google Drive offer paid options with additional capabilities, only the free versions are evaluated here.
The comparison is not meant to be exhaustive. However, if there is a category you would like to see compared, please contact us at firstname.lastname@example.org. Since web services evolve rapidly, the accuracy of the information is presented as-is. Therefore, it is recommended that particular attributes of interest be investigated in further detail before making a decision on which service to use.
Table 1: Comparison of cloud storage services. The green boxes indicate the most capable service(s) for that category. (Current as of 2016-05-10)
Basics (Basic information)
(for JHU users)
|Space||50 Gb||1 Tb||2 Gb||15 Gb|
|Max upload size of single file||10 Gb||2 Gb||10 Gb (web), Unlimited (Desktop/mobile)||5 Tb
(additional limits on uploaded files converted to Google Docs)
Ability to keep deleted or changed versions of files
(version retention policy not specified)
(Google Docs: unlimited, other formats: greater than 100 versions deleted after 30 days)
|Supported by JHU|
Management and tech support handled by JH IT
Syncing (Automated syncing between devices)
(for JHU users)
(requires encrypted system and application to JH IT)
(Android, iOS, Windows Phone. Use single sign-on option with
(Android, iOS, Windows Phone. Log in with
(Android, iOS, Windows Phone)
(Android, iOS, Windows Phone via unofficial client)
Collaboration (Capability to work with others)
(for JHU users)
Individuals external to JHU
(must have Box account)
(must have Microsoft Live ac-count, or an organizational account)
(must have Dropbox account)
(anonymous editing allowed)
|Real-time Document Editing||Yes|
(Box formats and Microsoft Office Online formats only)
(Microsoft Office Online for-mats only)
(functionality is upcoming: Dropbox Papers)
(Google formats only)
|Editable Sharing Permissions||Upload, Download, Preview, Get Link, Edit, Delete, Owner||View, Edit||View, Edit, Owner||View, Edit, Comment|
|Anonymous Link Sharing|
Others can view or edit content without logging into the service
|No||No||Yes (can view + add comments)||Yes (can view and edit Google docs)|
(for JHU users)
|Full text search||Yes|
(Office formats, plain text, text-only PDF, Google Docs/Sheets. 10000 character limit in all cases)
(Office formats only)
(Google Docs, plain text, first 100 pages of text-only PDFs, 10 pages of image-only PDFs)
|Approved by JHU for PII or PHI|
(if storing PII or PHI, JH-Box is the ONLY option allowed by JHU policy)
Increased security against unauthorized logins
(JHU supports MFA but it doesn’t seem to be implemented for Box)
(JHU supports MFA but it doesn’t seem to be implemented for OneDrive)
|Integration with Microsoft Office (Desktop)||Via Box Edit (edit any file using native app)||Full Office integration – open, save, collaborative editing|
(functionality varies between versions of office)
|None||Conversion from/to Office formats to Google Docs formats|
|Open Science Framework integration*|
Direct file upload / download
Best Practices for Cloud Services (Dos and Don’ts)
- Do sync only what you really need. It makes managing access and catching problems (e.g., missing or corrupted files) much easier.
- Do use the native formats of the service whenever possible. E.g., use a Google doc or an online Word document (in Google Drive and OneDrive respectively) instead of relying on syncing a traditional file. This allows you to take full advantage of the collaborative capabilities and versioning functionality of the format without having to deal with potential syncing problems. The main disadvantage is the limited functionality of those formats for complex tasks.
- Do use multi-factor authentication (MFA) – also known as two-factor authentication (2FA) – for increased resistance against unauthorized logins.
File organization and management:
- Do use a simple file structure without too many nested subfolders. It will make things easier to find and avoids potential problems when syncing to/from devices which may not support long file paths.
- Do create separate folders for separate groups which you will sharing files with. Place all files to be shared in these folders and avoid sharing individual files one-at-a-time.
- Do use descriptive file names. Despite the advanced search capabilities of some services, finding things may become needlessly difficult otherwise.
- Do use tags or the description field that some services provide to attach extra information to files.
- Don’t rely on cloud services as a primary backup solution (see the following section).
- Don’t assume that just because there were no errors during syncing, that the correct files are stored the way you intended. Sometimes sync clients behave in ways that are unexpected. Periodically use the web interface of the cloud service to check a couple files/folders to make sure the information is stored as intended.
- Don’t sync folders that are stored on the network drive as it may cause problems.
- Don’t sync everything. Syncing too many files (e.g., hundreds of thousands) may cause unpredictable or slow behavior for some sync clients. If there are many files, collect them together in a Zip or Tar archive first.
- Don’t make major directory structure changes outside of the service’s web interface. This way, the files stored in the cloud will accurately reflect your intended structure on all devices, avoiding potential unexpected behavior from the sync client.
Cloud Storage as a Backup Solution – Caution!
Although it is very tempting to use the automated syncing capabilities of the cloud services in Table 1 to back up files, such use should be avoided. If a file becomes corrupted (e.g., via a malware infection) or is accidentally deleted, the damaged or missing file will be automatically synced. Because there is a limited time window in which it can be recovered, data loss is possible. When only syncing a few files, it is easy to notice that something has gone wrong but when syncing thousands of files, the issue may go undetected until it’s too late. Apart from this, cloud storage solutions are not immune to bugs that may delete user files. For example, a bug in the Dropbox sync client deleted user files under certain circumstances (http://www.cnet.com/news/dropbox-fixes-file-deletion-bug-offers-year-of-free-service/).
In addition, the collaborative focus of cloud storage means that backups may inadvertently expose sensitive information unless steps are purposefully taken to secure this information (e.g., encrypting files before upload or setting appropriate permissions using the service’s web interface).
Although using cloud services as primary backup is generally discouraged, knowing the caveats of using them for this purpose means that they can be used for secondary backups of non-mission critical information. Some best practices to minimize risks include:
- Not using automated syncing – By manually uploading files, the risk of accidentally overwriting or corrupting the cloud copy of a file is minimized. Even if such an accident occurs, the likelihood of noticing the mistake is high due to the manual nature of the process. This means that the affected file can easily be recovered using the versioning capabilities of the service.
- Only the owner has the ability to write to the backup file/folder – All of the services mentioned have the capability of making files read only to others so permissions should be set appropriately. In essence, collaboration should not take place on files that are backups.
- Do not use the versioning capabilities of the services to back up files – this functionality is not intended to be used this way and may result in data loss depending on the type of file and the particular vendor used.
- Cloud storage should not be the only place where files are backed up – Storing files in the cloud for purposes other than syncing and collaboration is convenient, especially to satisfy the backup best practice of having offsite backups. However, in all cases, an external, offline backup is necessary to ensure that at least two backup copies of files exist. The offline backup should be considered the primary backup.
This work is licensed under a Creative Commons Attribution-NonCommercial 4.0 International License.