Cloud-based data storage certainly isn't perfected, although it's working very well for a great many enterprises. There's a lot of competition cropping up these days, and providers are elbowing each other in an effort to grab attention from potential customers, but don't be misled by claims that your data is always perfectly safe at all times.
At the IDG-IDC Cloud Leadership Conference in Santa Clara, Calif., on June 14 and 15, a snapshot poll of attendees was taken about some important issues. It didn't surprise anybody that security concerns were recorded as by far the No. 1 deterrent for enterprises thinking of -- but not acting on -- making the jump to cloud services.
Although cloud storage includes inherent vulnerabilities, they need not prevent a business user from taking advantage of its economies and flexibilities. Here, courtesy of CEO Andres Rodriguez of Nasuni -- former IT director of the New York Times -- are the five most important ways he considers to address data security in cloud-based storage infrastructures.
1. Data leakage: Many businesses that would benefit significantly from using cloud storage are holding back because of data leakage fear. The cloud is a multi-tenant environment, where resources are shared. It is also an outside party, with the potential to access a customer's data.
Sharing storage hardware and placing data in the hands of a vendor seem, intuitively, to be risky. Whether accidental, or due to a malicious hacker attack, data leakage would be a major security violation.
The best strategy is to assume from the start that the cloud vendor is compromised and send only encrypted files to the cloud. Use the strongest encryption that you can; anything less is not worthwhile.
Don't depend on the cloud provider or an intermediary to encrypt those files for you; in that case, they'll be able to decrypt them as well, and you'll have to rely on trust. With the cloud, all data and metadata should be encrypted at the edge before it leaves your premises. The only person to trust is yourself.
2. Cloud credentials: Even encrypted data can be vulnerable if your files are pooled in with those of another customer. Access to a given pool of storage is based on credentials, and if you are lumped together with another set of customers and share the same credentials, there is a risk that one of them could obtain those credentials and access your data.
They would not be able to decipher it, assuming it is encrypted, but they could delete the files. By securing your own unique credentials, however, your files will be separate. No one else will be able to log into your account and delete your data.
3. Snooping: Files can be vulnerable in the cloud, but there are also risks during data transmission. Strictly speaking, encrypted files do not need to be sent over a secure line; this amounts to double encryption. But it is best to assume the worst and guard against any measure of snooping by only sending and retrieving data over a secure line. This prevents against someone seeing cloud metadata. Data and metadata should be completely opaque on the wire and in the cloud.
Nothing -- even filenames and timestamps -- should be decipherable once it leaves your premises.
4. Key management: This has to be addressed properly, because if you botch key management, there is a risk that users will not want to activate the cryptography, which then compromises security. Key management should be so simple that users are not even aware of it:
--Encryption should be automatic. There should be no way to turn it off. This way, if there is no insecure mode, then there is no chance of someone accidentally sending unencrypted, vulnerable data to the cloud.
--Keys should also be securely escrowed and difficult to retrieve, so that no one can obtain that key to access your data. Ideally, you would escrow this key yourself, but Nasuni also offers customers secure key escrow.
5. Performance: A strong security strategy is a necessity, but it should not seriously impact performance. Encryption of data being sent to the cloud -- and decryption of files called back from the cloud -- should happen with little or no impact on the user experience. Ideally, it should all happen without the user noticing a thing.
Good advice all around, so thanks, Andres, for the contribution.
This might have been the most important article you read today.