国际资源子库 1.4专题资源 云计算与安全cloud computing security.pptVIP

* Mitigating Cloud Computing Availability Issues Risk analysts will tell you that when you confront a risk, you can try to eliminate the risk, you can mitigate/minimize the impact of the risk, or you can simply accept the risk. If you truly require non-stop availability, you can try using multiple cloud providers, or you could use public and private cloud nodes to improve redundancy. Some cloud computing services also offer service divided into multiple regions. By deploying infrastructure in multiple regions, isolation from single-region-only events (such as the power outage mentioned previously) can be obtained. Availability issues may also be able to be at least partially mitigated at the application level by things like local caching. Sometimes, though, it may simply make financial sense for you to just accept the risk of a rare and brief outage. (Remember, 99.99 availability== 52+ minutes downtime/yr) * Mitigating Data Loss Risks The risk of data loss (as in the T-Mobile Sidekick case) is an exception to the availability discussion on the preceding slide. Users may be able to tolerate an occasional service interrup-tion, but non-recoverable data losses can kill a business. Most cloud computing services use distributed and replicated global file systems which are designed to insure that hardware failures (or even loss of an entire data center) will not result in any permanent data loss, but I believe there is still value in doing a traditional off site backup of ones data, whether that data is in use by traditional servers or cloud computing servers. When looking for solutions, make sure you find ones that backs up data FROM the cloud (many backup solutions are meant to backup local data TO the cloud!) * Cloud Computing And Perimeter Security While Im not a huge fan of firewalls (as Ive previously discussed at the Spring 2008 I2MM in Cyberinfrastructure Architectures, Security and Advanced Applications, see ), at least some sites do find value in sheltering