​Visibility and control over SSL traffic in an era of HTTP/2.0

Author: Benn Alp, Solutions Architect of ANZ at F5 Networks

The emergence of attacks like POODLE and Heartbleed, has brought about renewed awareness about the importance of Secure Sockets Layer (SSL). At the enterprise level, SSL is what enables businesses to communicate securely with customers and partners. In order to keep SSL secure, businesses must have the visibility and control over SSL/ Transport Layer Security (TLS) traffic in the era of HTTP/2.0.

Why is the visibility and control of SSL/TLS traffic so important? Only a decade ago, SSL was reserved for financial institutions and some specific organisations like public sector agencies for the log-in pages of security-conscious web sites and services. Today this has been expanded to most web-based services, and is quickly becoming the de-facto protocol for communications. Based on industry research from Gartner, more than 50 percent of the world’s Internet traffic will be encrypted by the end of 2015[1].

TLS today offers a paradigm shift in which companies approach IT networks, as it offers communications security over a network server. The connection between client and server is private because symmetric cryptography is used to encrypt the data transmitted.

Securing SSL traffic in the enterprise

With the growing use of encrypted traffic, the traditional approach of depending on firewalls, intrusion protection systems and introduction detection systems has become obsolete, as these devices will be blindsided without knowing the data stream that is passing through it.

The solution to this problem is to have the initial SSL decryption take place at the security perimeter. However, most solutions available within the industry working at the security perimeter have not been designed or developed for the specific purpose of “SSL decryption”. While some businesses might have the capability to decrypt, they are not able to do so adequately. Many enterprises are also experiencing significant performance drop when they enable SSL decryption. This started a quest to explore “how” they should design their perimeter incorporating optimised handling of SSL traffic. Obviously, to maximise the efficacy of layer 7 security devices, the SSL decryption has to be executed near the security perimeter. Once the inbound SSL has been decrypted, the resulting requests can be analysed, modified, and steered.

SSL cipher

Cybercriminals often attack by using SSL to bypass security devices, knowing those devices are security loopholes. Malware hidden within SSL traffic could also easily bypass security platforms. Gartner also estimates that by 2017, more than 50 percent of attacks on the network of enterprises will use SSL to bypass security.

With the discovery of more vulnerable cipher suites that produces attacks like Heartbleed, BEAST, POODLE, and more, security administrators have a short time window to fix vulnerabilities. Having to manage hundreds or thousands of SSL fronting servers would take weeks to fix while they remain susceptible to attack from malicious sources.

It becomes a question of knowing what traffic to decrypt. If a business is serving content to users externally, it needs to use a device to offload SSL traffic from the server and then insert protection into the traffic flow. This will break the SSL, but in an intelligent way - you do not want to decrypt a banking session but you do for a Facebook session. Security needs to have the intelligence to understand where the traffic is going and then make a decision on whether it should be decrypted or left as it is.

Read more: Australians overconfident on security prowess despite surging toll of breaches

While SSL technically supports users to achieve security from point to point, it does not necessarily secure the entire traffic. SSL could be intercepted by hiding malware in the encrypted traffic without faking an SSL certificate. Therefore, the major focus for IT is to inspect the scams in real time and mitigate them immediately.

The is security proxy architectures that can enable conversion and decryption seamlessly, while implementing separated connections for both internal and external communications. Architectures can also provide the ability to terminate SSL sessions to make it easier for IT to use encryption for traffic in the outside and utilise the old HTTP if they need to. All connections will be possible and IT will not have to break and replace the whole web application infrastructure to enjoy the benefits of HTTP/2.0.

Keep the performance while managing TLS configuration

With the adoption of stronger key lengths, from 1024 bits to 2048 bits, computational requirements have increased exponentially from four to eight times that of the previous. This decreases performance on the existing infrastructure that has to handle a similar traffic size. Having specialised hardware together with a hardware accelerator will alleviate the need to upgrade CPU resources across the entire server farm.

With PCI DSS 3.1 mandating companies to stop using SSL and TLS 1.0 by 30 June 2016 in favour of more recent TLS implementations, companies will need to comply with all devices having a centralized SSL management point that will allow issues to be rectified within minutes across the entire environment.

Whether or not the world wants to admit it, the stakes around Internet Security have been raised into the same spheres as human rights, freedom of speech and free commerce. It is time for enterprises to increase overall security posture to protect their apps, data and most of all their last line of defense, the SSL.


Read more: ​Re-used crypto keys expose millions of devices to attack

[1] Gartner “Responding to New SSL Cybersecurity threats” https://www.venafi.com/collateral/wp/gartner-responding-to-new-ssl-cybersecurity-threats

Join the CSO newsletter!

Error: Please check your email address.

Tags F5 NetworksSSL traffic​VisibilityCSO Australia

More about FacebookGartnerTransport

Show Comments