Users need to become better at asking cloud providers questions about the finer points of availability and vulnerability management in contracts, according to a new guide from the European Network and Information Security Agency (ENISA).
ENISA's goal is to improve the public sector's understanding of the security of cloud services and the potential indicators and methods that can be used during service delivery, the authors said.
Procuring and managing service contracts for cloud services is an increasingly important task for IT staff.
"Organisations have started switching from running systems internally to outsourcing and using cloud services. So the skills and focus of IT staff have to change," Marnix Dekker, who co-authored the report.
An integral part of that work is to specify security requirements up front, but even more important to be able to monitor and verify whether these security requirements are being met throughout the lifetime of the contract, according to the guide.
"You need to be sure that the solution you are buying fits your security requirements," said Dekker.
So far, a lot of the focus has been on ensuring that the right security measures are in place, but not as much on how they perform when the cloud service is up and running, according to Dekker.
"A good example is patch management. In the request for proposal you ask for patch management to be in place. As a customer you also want to know on an ongoing basis whether patches have been applied, but that part is often left unaddressed," said Dekker.
The guide covers eight different parameters that IT staff need to be on top of. Among the most important ones are service availability, incident response, and technical compliance and vulnerability management, according to Dekker.
When it comes to service availability - which isn't just a security requirement, but a business requirement - people are still having trouble specifying how it should be defined and reported on.
"What you often see is a very generic statement that the service should be up and running, without thinking about what functions should be up and running," he said.
It may take 15 minutes to send an email, but the cloud service provider can still say that the service is not down, according to Dekker. That's why it's important to think beyond a service just being available, and how that should be measured, he said.
Incident response, which could be related to availability or security, is also an important one.
"It is difficult to write a contact on how quick incidents should be resolved, because they can be very complicated to address. But for users it is important to get data on how fast incidents were or were not resolved " said Dekker.
Even if there are issues that need to considered when moving to the cloud, ENISA sees it as an opportunity to make systems more secure.
"Because of the economics of scale, cloud providers can hire experts and have full time security personnel," said Dekker.
But only if it is done right.
"When users are more mature and ask the right questions the best cloud providers will be able to answer them, and security will improve," said Dekker.