IBM Says Cloud Computing Security is Less Susceptible to Attacks If Done Right

Cloud security requirements are different from what traditional security models can provide

In a newly published eBook, IBM has outlined exactly why cloud infrastructure is safer than traditional datacenters. The eBook is called: Should We Fear the Cloud? It may be the key to security, and covers various risk factors as they apply to cloud environments as opposed to company-owned datacenters.

According to the IBM report, the average cost of a data breach has gone up by 15 percent in the past one year and is now estimated at $3.5 million. The report goes on to say that data breaches can cause companies to lose customers, and this phenomenon is more acute in financial services, healthcare and pharmaceuticals.

In a survey of 250 senior IT and business decision makers in the UK, only 2 percent said they had experienced a cloud-related security breach.

The top five security threats today are data breaches, data loss, service traffic hijacking, insecure interface and API, and denial-of-service attacks.

IBM shows that cloud-based security approaches are often more secure than traditional methods used in in-house datacenters.

The consensus is that speed and skill are critical in the event of a data breach, and that working with a consultant to deploy a unified data breach response system is one solution to mitigate this risk.

Where data loss is concerned, IBM suggests that endpoint security for devices offers significant protection against the risk of data being lost or stolen.

Service traffic hijacking is also one of the top security threats, and it’s simply because developers typically tend to trust that users will never perform malicious actions. A big mistake, and the high frequency of these attacks (cross-site scripting or XSS attacks) shows how widespread this problem really is. IBM suggests that contextual output coding or escaping is usually the primary defense against XSS attacks, and recommends the use of a security encoding library.

Interfaces and application programming interfaces, or APIs, not being secure means an additional vulnerability in any system, including cloud. IBM highlights the need for a secure cloud service provider, and recommends secure access using multilevel security (MLS) and mandatory access control (MAC).

The last of these security threats is one that many of us are now familiar with – distributed denial-of-service attacks, or DDoS attacks. Though cloud servers are equally susceptible to such attacks, IBM says that constant monitoring and early warnings can help mitigate such events, and virtual machines and bare metal servers that have been attacked and hijacked as ‘zombies’ should have their DDoS attacks blocked and suspended.

In short, IBM says that traditional security models don’t necessarily work on the cloud even though the attack methods may be traditional. Cloud providers, in the past, have used traditional perimeter-based static controls like firewalls and intrusion protection systems (IPS), as well as additional defense layers, under the assumption that more integrated layers mean better security.

But IBM says that this leads to other problems, such as numerous security controls leading to a never-ending stream of alerts that put additional overhead pressure on security management. Another ‘fallout’ effect of traditional security models is that the current sophistication of attacks can more easily bypass static security measures.

The report goes on to explain new paradigms of security that are essential to cloud environments, and recommends an industry-specific approach to creating new security models. You can view the full eBook report here.

Thanks for reading our work! We invite you to check out our Essentials of Cloud Computing page, which covers the basics of cloud computing, its components, various deployment models, historical, current and forecast data for the cloud computing industry, and even a glossary of cloud computing terms.