|
George Reese (author of the new book Cloud Application Architectures: Building Applications and Infrastructure in the Cloud) is talking at Gluecon about securing cloud infrastructures. Two recent surveys found “security” was the number one concern of companies considering a move to the cloud. George says the key to making customers comfortable with cloud security is transparency.
Without security:
So… demand transparency. That ought to be a critical part of deciding what cloud infrastructure to use. Control isn’t the real issue: people don’t build their own microchips. People give up control when they are comfortable. Control can be an issue with out transparency.
There are different security concerns at different levels. Are we talking about the infrastructure level? The platform level? The service level?
Some key control issues:
As dependency on cloud services grows, identity management becomes a greater challenge. If you use Constant Contact for email, Salesforce for CRM, someone else for accounting, etc. do you have to have different IDs for each? What about password policies? What about divergent security policies? Access to the infrastructure itself is often difficult using traditional ID management solutions.
Host intrusion detection systems (HIDS) work fine on cloud infrastructure, but are hard to do at higher levels of the stack. Network intrusion detection systems (NIDS) are impossible to do at most providers. The traditional notion of “perimeter” is not necessarily available in the cloud.
Audit controls at the infrastructure layer are often weak. They don’t exist at all at the platform or service levels. There aren’t many options for write-once data stores. Getting PCI level 1 compliance requires combining the cloud service (like AWS) with a specific vendor who provides a PCI compliant service (like Aria).
On the other hand, the cloud makes it possible to have a proven disaster recovery plan. Cloud computing can alter the dynamics of application availability and the ability to develop solid disaster recovery procedures. For example, you can do daily automated testing of your procedures and have a fully redundant capability at nearly zero cost.
Sponsored byCSC
Sponsored byWhoisXML API
Sponsored byIPv4.Global
Sponsored byRadix
Sponsored byVerisign
Sponsored byDNIB.com
Sponsored byVerisign