Every day, attackers conspire to take down applications and steal data, leaving data centre infrastructure in the crosshairs.
Storing an organisation’s most valuable and most visible assets – its web, DNS, database and email servers – data centres have become the number one target of cyber criminals, hacktivists and state-sponsored attackers.
Whether seeking financial gain, competitive intelligence or notoriety, attackers are carrying out their assaults using a range of weapons. The top five most dangerous threats to a data centre are:
- DDoS attacks
- Web application attacks
- DNS infrastructure: attack target and collateral damage
- SSL-induced security blind spots
- Brute force and weak authentication
To counter these threats, organisations need a solution that can lock down their data centres. Otherwise they risk a high-profile data breach, downtime or even brand damage.
Servers are a prime target for distributed denial of service (DDoS) attacks aimed at disrupting and disabling essential internet services.
While web servers have been at the receiving end of DDoS attacks for years, attackers are now exploiting web application vulnerabilities to turn web servers into ‘bots’. They use these captive servers to attack other websites.
By leveraging web, DNS and NTP servers, attackers can amplify the size and the strength of DDoS attacks. While servers will never replace traditional PC-based botnets, their greater compute capacity and bandwidth enable them to carry out destructive attacks - one server could equal the attack power of hundreds of PCs.
With more and more DDoS attacks launched from servers, it’s not surprising that the size of attacks has grown sharply. Between 2011 and 2013, the average size of DDoS attacks escalated surged from 4.7 to 10 Gbps.
Worse, there has been the staggering increase in the average packets per second in typical DDoS attacks; attack rates skyrocketed 1,850% to 7.8 Mpps between 2011 and 2013. At the current trajectory, DDoS attacks could reach 75 Mpps in 2015 - powerful enough to incapacitate most standard networking equipment.
DDoS for hire services, often called ‘booters’, have mushroomed too. Many advertise their capabilities in YouTube videos and forum posts. While some masquerade as ‘stress testing’ services, many boldly claim to ‘take enemies offline’ or ‘eliminate competitors’. Such services enable virtually any individual or organisation to execute a DDoS attack.
Web application attacks
Cyber criminals also launch web attacks like SQL injection, cross-site scripting (XSS) and cross-site request forgery (CSRF), trying to break into applications and steal data for profit. Increasingly, attackers target vulnerable web servers and install malicious code in order to transform them into DDoS attack sources.
Some 98% of all applications currently have or have had vulnerabilities, and the median number of vulnerabilities per application was 20 in 2014, according to a 2015 Trustwave Global Security Report.
Today’s most dangerous application threats, like SQL injection and cross-site scripting, aren’t new but they are still easy to perform and lethally effective. Tools like the Havij SQL injection tool enable hackers to automate their attack processes and quickly exploit vulnerabilities.
The recent wave of web attacks on CMS applications has also revealed a gaping hole in the strategy to lock down applications by writing secure code. Because CMS applications are usually developed by third parties, organisations can’t rely on the protection of secure coding. In 2013, 35% of all breaches were caused by web attacks. More than ever, organisations need a proactive defence to block web attacks and ‘virtually patch’ vulnerabilities.
DNS servers have become a top attack target for two reasons. First, taking DNS servers offline is an easy way for attackers to keep thousands or millions of Internet subscribers from accessing the Internet. If attackers incapacitate an ISP’s DNS servers, they can prevent the ISP’s subscribers from resolving domain names, visiting websites, sending email and using other vital Internet services.
Secondly, attackers can exploit DNS servers to amplify DDoS attacks. In DNS reflection attacks, attackers spoof the IP address of their real attack target. They send queries that instruct the DNS server to recursively query many DNS servers or to send large responses to the victim. As a result, powerful DNS servers drown the victim’s network with DNS traffic. Even when DNS servers are not the ultimate target of the attack, they can still suffer downtime and outages as the result of a DNS reflection attack.
SSL-induced blind spots
To prevent the continuous stream of malware and intrusions in their networks, enterprises need to inspect incoming and outgoing traffic for threats. Unfortunately, attackers are increasingly turning to encryption to evade detection.
With more and more applications supporting SSL – more than 40%of applications can use SSL or change ports – SSL encryption represents an enormous crater that malicious actors can exploit.
While many firewalls, intrusion prevention and threat prevention products can decrypt SSL traffic, they can’t keep pace with growing SSL encryption demands. The transition from 1024- to 2048-bit SSL keys has burdened security devices because 2048-bit certificates require approximately 6.3 times more processing power to decrypt. With SSL certificate key lengths continuing to increase, many security devices are collapsing under increased decryption demands.
For end-to-end security, organisations need to inspect outbound SSL traffic originating from internal users, and inbound SSL traffic originating from external users to corporate-owned application servers to eliminate the blind spot in corporate defences.
NSS Labs found that eight leading next-generation firewall vendors experienced significant performance degradation when decrypting 2048-bit encrypted traffic. NSS Labs asserted that it had ‘concerns for the viability of SSL inspection in enterprise networks without the use of dedicated SSL decryption devices’.
Clearly organisations need a high-powered solution to intercept and decrypt SSL traffic, offloading intensive SSL processing from security devices and servers.
Brute force and weak authentication
Applications often use authentication to verify users’ identity, allowing application owners to restrict access to authorised users and customise content based on user identity. Unfortunately, many enforce only single-factor, password-based authentication. This exposes them to a host of threats, from simple password guessing and stolen credentials to automated brute force attacks from password-cracking tools.
Many users select the same password for multiple accounts, so when one is compromised, all others are at risk. Within hours of a breach, hackers will crack stolen password lists, even password hashes, and use them to break into other online accounts.
Two-factor authentication can drastically reduce the risk of password cracking. Combining passwords with out-of-band authentication such as SMS messages to mobile devices or with hardware or software tokens greatly decreases the risk of brute force or password cracking. In addition, user context, such as a user’s browser and operating system or a user’s geographic location, can help to identify fraudulent activity. Application owners can build advanced rules to identify high-risk users or password-cracking tools, to safeguard user accounts.
Simply rolling out and managing authentication across many different web applications can be daunting, while setting up client authentication schemes for dozens of applications entails costly and time-consuming development work. So organisations need an integrated solution that can centrally manage authentication services and block users with repeated failed login attempts.
ADCs offer multiple protection
To shield data centre infrastructure from attack, organisations need a solution that can mitigate a multitude of threat vectors and still deliver unmatched performance.
Application delivery controllers (ADCs) can help organisations to safeguard their data centre infrastructure. Deployed in the heart of the data centre, ADCs can block attacks, intercept and inspect encrypted traffic and prevent unauthorised access to applications.
Next-generation ADCs offer the following defences to shield data centre infrastructure from emerging threats: DDoS protection, Web application firewall (WAF), DNS application firewall (DAF), SSL insight and SSL Offload, and application access management for authentication.
Organisations should evaluate the security features of ADCs carefully to make sure they can mitigate data centre risks effectively. They should seek a product line that helps to protect servers and applications from data centre risks, while still providing unmatched application performance. An ADC that includes a comprehensive set of security features at no additional cost can be a bonus.
By Greg Barnes, Managing Director, ANZ, A10 Networks