Malicious software or malware is a security nightmare. One can agree with the statement that malware in the internet is a growing epidemic and is costing industry billions of dollars each year.

According to Microsoft Malware "is short for malicious software and is typically used as a catch-all term to refer to any software designed to cause damage to a single computer, server, or computer network, whether it's a virus, spyware, et al."

Although the growth of malware may be flattening, its design sophistication is not. Per Kaspersky labs approximately 300,000 new malware programs were discovered per day; last year.

Graham Cluley of Sophos Labs stated Sophos "sees in excess of 50,000 new samples daily the vast majority of which are focused on making money."

These days your typical malware agent is designed to conceal itself in an intelligent and proactive/reactive manner from standard AV software.

As we know there are several ways malware can infect a machine e.g. drive by download, infected USB, spoofing, fake anti-virus security products, search engine optimisation (SEO) by drawing internet users to compromised and legitimate websites and infecting their machines as a result etc.

In order to maintain its residence on an infected system/s malware has to either

a)  Adapt legitimate data or program, or programs within the client device, in order to gain control after the audit process has completed or

b) Remain active in RAM or swap space

How does this relate to the cloud? Assuming a capitalizing on the much hyped - scalability and flexibility; a malware agent can potentially forward relevant data into a compromised cloud system, thus capitalizing on the cloud's scalable computational and storage properties.

It can then use this access to "infinite computing power and storage" to compute values needed by the verification authority on the infected  system thereby avoiding detection and possibly moving undetected throughout any array of networks; wreaking undetected havoc as it processes data.

But the cloud is a virtual environment; and virtualization is highly dependent on a hypervisor which is easily verifiable and well protected...right?

Not so much: this premise was in part due to the hypervisor or Virtual Machine Monitor (VMM), initially consisting of a comparatively small code base which had limited interaction with the external world.

However there are known attacks on the hypervisor which could possibly impact the Cloud Computing Ecosystem: two of which are

1. Hyper-jacking: Where an attacker inserts a rogue hypervisor; taking advantage of an assumption that there is a level of difficult for any OS running on the hypervisor to detect an attacker's insertion. This based on the level of privilege at which the hypervisor runs. An attack of this nature can potentially give control of any virtual machine running on the physical server to the attacker e.g. Joanna Rutkowska's Blue Pill rootkit.

2. VM Escape: If a virtual system is compromised by an attacker and via a set software system is able to totally bypass the hypervisor or virtual layer, it is quite possible to get access to the host machine and by this root privileges. For instance a vulnerability vmnat.exe was identified in VMware Workstation, which could be exploited by remote attackers to execute arbitrary commands.

The exploit that enables VM Escape can allow an attack on a specific virtual server to escalate from the virtual server to the underlying user space. An attacker can then gain control of the underlying hypervisor.

Of importance is although such exploits can be used under certain circumstances to cause the host environment to execute arbitrary code, this can only occur with weak security configurations e.g. weak code, unpatched systems.

In terms of the hypervisor code we are seeing a growth in its volume with time e.g. recent attacks on Xen's showed that its code and data can be modified at runtime, allowing backdoor functionality. All known issues were patched however with the growth in the skills and resources within the "hacking enterprises", there should be some concern and planning for risk mitigation within the environment.

A recent study stated that VMware ESX 3.x bare-metal hypervisor exhibited at least 165 known vulnerabilities.

In light of this, a potential cloud customer should get a good working idea of their vendor's security policies, continuity and recovery protocols to ensure the confidentiality, integrity and availability of their data.

References

1. Virtualization Security. Edward Ray, Eugene Schultz. s.l. : ACM -The Association for Computing Machinery, 2004. 1-58113-000-0/00/0004.
2. Tim Shelton. [VulnWatch] [ACSSEC-2005-11-25-0VMWare Workstation 5.5.0
3. Carl Gebhardt, Allan Tomlinson. Security consideration for Virtualization. Egham : Depart of Mathematics Royal Holloway University
4. Paul Arcenaux. Securing the virtual data center. www.hp.com. [Online] Hewlitt Packard (HP), May 2010.
5. K.-K. Choo. Organised crime groups in cyberspace: a typology. Trends in Organized Crime, 11(3):270{295, 2008
6. Network Security April 2010 pgs 8-10

• Cloud Jackin, Hacking the Cloud
• Securing & Hacking the Cloud
• Panda Security Upgrades Cloud-Based Anti-Malware Service