52 Things: Number 49: Describe the basic ideas behind IPSec and TLS.

This is the latest in a series of blog posts to address the list of '52 Things Every PhD Student Should Know To Do Cryptography': a set of questions compiled to give PhD candidates a sense of what they should know by the end of their first year. This week we discuss the basic ideas behind IPSec and TLS.

Internet Protocol Security (IPsec) and Transport Layer Security (TLS) both aim to create a secure communication channel between two parties over an insecure network. In general, both use some mechanism to establish a private session key (either pre-shared or via a key negotiation protocol) and use symmetric key cryptography for the bulk of the communication. There are some further details with regards to authentication but I'll skip over that. Although these two ultimately have similar goals, they differ considerably in their implementation.

IPSec sits on the network layer of the OSI model and aims to provide integrity, authenticity and confidentiality between two end points. As it sits on network layer, it blindly encrypts, MACs and packages up the data from the above layers before sending it down the line. This effectively creates a virtual network link between the two end-points without the need to ensure the end-point application has secured the data appropriately. This is often deployed for enterprise VPN solutions as it is a fast solution for remote access to an enterprise network. The downside however is that once a connection is up, it is tricky to restrict applications from using the connection once it is up.

TLS on the other hand establishes a secure connection at the application layer of the OSI model. We see TLS heavily used for securing web protocols such as HTTPS, STARTTLS etc. and as a consequence, each connection/application will negotiate/set up a secure connection independently. From a security perspective, this is quite attractive as a single compromised channel *should* have no bearing on the remaining channels. Whilst TLS can be viewed as a more flexible approach, it does incur some overhead over IPSec for a large number of connections between two nodes.

It's easy to get into very fine details but I think that should cover the 'basic' ideas of the two.