Secure Cloud Computing with a Virtualized Network Infrastructure
Despite the rapid development in the ﬁeld of cloud computing, security is still one of the major hurdles to cloud computing adoption. Most cloud services (e.g. Amazon EC2) are offered at low cost without much protection to users. At the other end of the spectrum, highly secured cloud services (e.g. Google “government cloud”) are offered at much higher cost by using isolated hardware, facility, and administrators with security clearance. In this paper, we explore the “middle ground”, where users can still share physical hardware resource, but user networks are isolated and accesses are controlled in the way similar to that in enterprise networks. We believe this covers the need for most enterprise and individual users. We propose an architecture that takes advantage of network virtualization and centralized controller. This architecture overcomes scalability limitations of prior solutions based on VLANs, and enables users to customize security policy settings the same way they control their on-site network.
Despite the rapid development in the ﬁeld of cloud computing, security is still one of the major obstacles to cloud computing adoption. To ease the concerns of IT managers, it is critical to ensure data privacy and integrity in the cloud at a level that is at least comparable to that in current enterprise networks. However, the current cloud computing services fall in short on isolating computing resources and networks between customers. This is not surprising because the success of cloud computing depends on economy of large scales. It is essential for cloud service providers to take advantage of resource sharing and multiplexing among customers. Virtual machines of diﬀerent customers may reside on the same physical machine, and their data packets may share the same LAN. Such lack of isolation brings security risks to users. For example, has shown that it is possible for a hacker to conduct attacks towards another Amazon EC2 user who shares hardware resources with the hacker in the cloud. On the other end of the spectrum, google has proposed “government cloud”, which creates entirely separate hardware, software, and administrators (with appropriate background checks) for special customers. While such cloud service can be very secure, it is also very expensive — almost like building a separate data center for each customer. In this paper, we explore the “middle ground”, where users can still share physical hardware resource, but user networks are isolated and accesses are controlled in the way similar to that in enterprise networks. We believe this covers the need for most enterprise and individual users. More speciﬁcally, we propose a new data center architecture with following properties:
• Isolation. The architecture provides effective isolation between diﬀerent customer networks. This includes supporting their private IP address spaces,
which may potentially be overlapping, and isolating their traﬃc. Resource allocation should be
managed so that customers cannot impact each
other’s resource usage in an uncontrolled manner.
• Transparency. The underlying data center infrastructure and hardware should be transparent to
the customers. Each customer should have a logical view of its own network, independent of the
actual implementation. This simpliﬁes the administration for the customer and improves security.
• Location independence. The virtual machines (VM)
and networks of customers should be “location independent”, i.e., can be physically allocated anywhere in the data center. This can greatly improve
resource utilization and simplify provisioning.
• Easy policy control. Each customer may have its
own policy and security requirements. The architecture should allow customers to conﬁgure their
individual policy settings on the ﬂy, and enforce
such settings in the network.
• Scalability. The number of customers that can
be supported should be restricted only by the resources available in the data center, not by design
• Low cost. The solution must mostly rely on oﬀ-
the-shelf devices, so that new investment for cloud
service providers can be reduced.