diff --git a/index.rst b/index.rst
index a35b333..f7b723f 100644
--- a/index.rst
+++ b/index.rst
@@ -18,6 +18,7 @@ Peterson and Davie
    :caption: Table of Contents
    :glob:
 
+   preface.rst
    intro.rst
    principles.rst
    crypto.rst
diff --git a/preface.rst b/preface.rst
new file mode 100644
index 0000000..91c7702
--- /dev/null
+++ b/preface.rst
@@ -0,0 +1,94 @@
+Preface
+========
+
+The field of network security is roughly as old as networking
+itself. Conventional wisdom is that the original Internet was built
+without security features and we have been dealing with the effects of
+those design decisions ever since. And it is not that the Internet's
+designers, implementers and architects were unaware of security
+concerns; many of them were directly involved in developing security
+technologies in early operating systems. But building a
+high-performance network that could scale to global proportions and
+accommodate the heterogeneous set of technologies that existed (and
+those still to come) presented more than enough challenges. To quote
+Internet pioneer Vint Cerf: *"getting this thing to work at all was
+non-trivial.”*
+
+Adding security capabilities to the Internet and the systems connected
+to it has been a multi-decade effort. While there have been some
+notable successes, such as widespread adoption of Transport Layer
+Security (TLS), vulnerabilities continue to be found and exploited. In
+part this is because security is a "negative goal": we are always
+trying to ensure that there are no more cracks in our defenses, and it
+is fundamentally impossible to prove that we are "done". So security
+continues to evolve as new weaknesses are exploited and new mechanisms
+deployed to prevent further exploits.
+
+We have been educating ourselves about network security since we wrote
+the first edition of *Computer Networks: A Systems Approach*
+in 1995. There is an adage in security circles that no-one should
+write their own cryptography code because it is so hard to get right,
+and something similar might be said about trying to write a security
+book. It is very easy to make mistakes, especially if you are not
+deeply immersed in the security world and the "look for every possible
+weakness" mentality. We've had to make a few corrections over the
+years to our security section in the big textbook.
+
+So why did we decide to write the current book? First, we saw an
+opportunity to write about security in a way that would make sense to
+a networking person. Also, we saw an opportunity to take more of a
+systems approach to the topic. While we always try to
+take a system-level view in everything we write, it's easy with
+security to get bogged down in the details of individual components
+such as cryptographic algorithms without really tackling the systems
+issues. Cryptography is cool and interesting (in our view at least) but
+it isn't really the main thing to focus on if you are building secure
+systems. So while we do explain the basics of cryptography here, it's
+not the focus. We're aiming to explain how a system that comprises
+many moving parts, both in the network and the end-system, can be made
+secure.
+
+This question of focus caused us to examine how much we wanted to say
+about end-system security. There are entire books to be written on
+operating system security, processor architecture bugs such as spectre
+and meltdown, and preventing malware on end-systems. We made a
+conscious decision to draw a line around the network and focus there,
+recognizing that, just as TCP is an important network protocol that
+runs in end-systems, protocols like HTTPS and TLS need to be covered
+in a book on network security. In fact TLS provides an excellent case
+study in the system-level issues that come into play when you try to
+secure traffic that flows between end-systems over the Internet, and
+we devote an entire chapter to it.
+
+With the never-ending set of threats and vulnerabilities that need to
+be dealt with, it is all too easy to start thinking of security as
+just a collection of point solutions to the problems that have been
+identified so far. But in fact there is a well-established set of
+principles that have been identified and written down by pioneers in
+the field, such as the principle of least privilege, defense in depth,
+and so on. We have dedicated a chapter to exploring some of the most
+widely accepted principles, and then we see them applied repeatedly
+throughout the book. Perimeter firewalls, for example, can be a part of a defense
+in depth strategy, while *distributed* firewalls have been proposed
+as a way to apply least privilege to datacenter networks.
+
+Inevitably there are security technologies and types of attack that we
+have not covered in this book. What we have tried to do is to give the
+reader a sense of the principles at work and the building blocks that
+exist to build secure systems, and then illustrate how the blocks can
+be assembled to tackle particular problems ranging from securing Web
+traffic to securing the routing system that underpins the
+Internet. Security is a never-ending set of challenges and there will
+always be new threats and solutions. This book aims to lay the
+foundation so you can understand the network security landscape as it
+continues to evolve.
+
+
+
+Acknowledgements
+----------------
+
+Brad Karp
+Cecilia Testart
+Motonori Shindo
+Nick Feamster