William Cheswick

Bio: William Cheswick is an academic researcher from Bell Labs. The author has contributed to research in topics: The Internet & Encryption. The author has an hindex of 7, co-authored 7 publications receiving 1623 citations.

Firewalls and Internet Security: Repelling the Wily Hacker

Abstract: From the Book: But after a time, as Frodo did not show any sign of writing a book on the spot, the hobbits returned to their questions about doings in the Shire. Lord of the Rings —J.R.R. TOLKIEN The first printing of the First Edition appeared at the Las Vegas Interop in May, 1994. At that same show appeared the first of many commercial firewall products. In many ways, the field has matured since then: You can buy a decent firewall off the shelf from many vendors. The problem of deploying that firewall in a secure and useful manner remains. We have studied many Internet access arrangements in which the only secure component was the firewall itself—it was easily bypassed by attackers going after the “protected” inside machines. Before the trivestiture of AT&T/Lucent/NCR, there were over 300,000 hosts behind at least six firewalls, plus special access arrangements with some 200 business partners. Our first edition did not discuss the massive sniffing attacks discovered in the spring of 1994. Sniffers had been running on important Internet Service Provider (ISP) machines for months—machines that had access to a major percentage of the ISP’s packet flow. By some estimates, these sniffers captured over a million host name/user name/password sets from passing telnet, ftp, and rlogin sessions. There were also reports of increased hacker activity on military sites. It’s obvious what must have happened: If you are a hacker with a million passwords in your pocket, you are going to look for the most interesting targets, and .mil certainly qualifies. Since the First Edition, we have been slowlylosing the Internet arms race. The hackers have developed and deployed tools for attacks we had been anticipating for years. IP spoofing Shimomura, 1996 and TCP hijacking are now quite common, according to the Computer Emergency Response Team (CERT). ISPs report that attacks on the Internet’s infrastructure are increasing. There was one attack we chose not to include in the First Edition: the SYN-flooding denial-of- service attack that seemed to be unstoppable. Of course, the Bad Guys learned about the attack anyway, making us regret that we had deleted that paragraph in the first place. We still believe that it is better to disseminate this information, informing saints and sinners at the same time. The saints need all the help they can get, and the sinners have their own channels of communication.Crystal Ball or Bowling Ball?The first edition made a number of predictions, explicitly or implicitly. Was our foresight accurate? Our biggest failure was neglecting to foresee how successful the Internet would become. We barely mentioned the Web and declined a suggestion to use some weird syntax when listing software resources. The syntax, of course, was the URL... Concomitant with the growth of the Web, the patterns of Internet connectivity vastly increased. We assumed that a company would have only a few external connections—few enough that they’d be easy to keep track of, and to firewall. Today’s spaghetti topology was a surprise. We didn’t realize that PCs would become Internet clients as soon as they did. We did, however, warn that as personal machines became more capable, they’d become more vulnerable. Experience has proved us very correct on that point. We did anticipate high-speed home connections, though we spoke of ISDN, rather than cable modems or DSL. (We had high-speed connectivity even then, though it was slow by today’s standards.) We also warned of issues posed by home LANs, and we warned about the problems caused by roaming laptops. We were overly optimistic about the deployment of IPv6 (which was called IPng back then, as the choice hadn’t been finalized). It still hasn’t been deployed, and its future is still somewhat uncertain. We were correct, though, about the most fundamental point we made: Buggy host software is a major security issue. In fact, we called it the “fundamental theorem of firewalls”: Most hosts cannot meet our requirements: they run too many programs that are too large. Therefore, the only solution is to isolate them behind a firewall if you wish to run any programs at all. If anything, we were too conservative.Our ApproachThis book is nearly a complete rewrite of the first edition. The approach is different, and so are many of the technical details. Most people don’t build their own firewalls anymore. There are far more Internet users, and the economic stakes are higher. The Internet is a factor in warfare. The field of study is also much larger—there is too much to cover in a single book. One reviewer suggested that Chapters 2 and 3 could be a six-volume set. (They were originally one mammoth chapter.) Our goal, as always, is to teach an approach to security. We took far too long to write this edition, but one of the reasons why the first edition survived as long as it did was that we concentrated on the concepts, rather than details specific to a particular product at a particular time. The right frame of mind goes a long way toward understanding security issues and making reasonable security decisions. We’ve tried to include anecdotes, stories, and comments to make our points. Some complain that our approach is too academic, or too UNIX-centric, that we are too idealistic, and don’t describe many of the most common computing tools. We are trying to teach attitudes here more than specific bits and bytes. Most people have hideously poor computing habits and network hygiene. We try to use a safer world ourselves, and are trying to convey how we think it should be. The chapter outline follows, but we want to emphasize the following: It is OK to skip the hard parts. If we dive into detail that is not useful to you, feel free to move on. The introduction covers the overall philosophy of security, with a variety of time-tested maxims. As in the first edition, Chapter 2 discusses most of the important protocols, from a security point of view. We moved material about higher-layer protocols to Chapter 3. The Web merits a chapter of its own. The next part discusses the threats we are dealing with: the kinds of attacks in Chapter 5, and some of the tools and techniques used to attack hosts and networks in Chapter 6. Part III covers some of the tools and techniques we can use to make our networking world safer. We cover authentication tools in Chapter 7, and safer network servicing software in Chapter 8. Part IV covers firewalls and virtual private networks (VPNs). Chapter 9 introduces various types of firewalls and filtering techniques, and Chapter 10 summarizes some reasonable policies for filtering some of the more essential services discussed in Chapter 2. If you don’t find advice about filtering a service you like, we probably think it is too dangerous (refer to Chapter 2). Chapter 11 covers a lot of the deep details of firewalls, including their configuration, administration, and design. It is certainly not a complete discussion of the subject, but should give readers a good start. VPN tunnels, including holes through firewalls, are covered in some detail in Chapter 12. There is more detail in Chapter 18. In Part V, we apply these tools and lessons to organizations. Chapter 13 examines the problems and practices on modern intranets. See Chapter 15 for information about deploying a hacking-resistant host, which is useful in any part of an intranet. Though we don’t especially like intrusion detection systems (IDSs) very much, they do play a role in security, and are discussed in Chapter 15. The last part offers a couple of stories and some further details. The Berferd chapter is largely unchanged, and we have added “The Taking of Clark,” a real-life story about a minor break-in that taught useful lessons. Chapter 18 discusses secure communications over insecure networks, in quite some detail. For even further detail, Appendix A has a short introduction to cryptography. The conclusion offers some predictions by the authors, with justifications. If the predictions are wrong, perhaps the justifications will be instructive. (We don’t have a great track record as prophets.) Appendix B provides a number of resources for keeping up in this rapidly changing field.Errata and UpdatesEveryone and every thing seems to have a Web site these days; this book is no exception. Our “official” Web site is . We’ll post an errata list there; we’ll also keep an up-to-date list of other useful Web resources. If you find any errors—we hope there aren’t many—please let us know via e-mail at .AcknowledgmentsFor many kindnesses, we’d like to thank Joe Bigler, Steve “Hollywood” Branigan, Hal Burch, Brian Clapper, David Crocker, Tom Dow, Phil Edwards and the Internet Public Library, Anja Feldmann, Karen Gettman, Brian Kernighan, David Korman, Tom Limoncelli, Norma Loquendi, Cat Okita, Robert Oliver, Vern Paxson, Marcus Ranum, Eric Rescorla, Guido van Rooij, Luann Rouff (a most excellent copy editor), Abba Rubin, Peter Salus, Glenn Sieb, Karl Siil (we’ll always have Boston), Irina Strizhevskaya, Rob Thomas, Win Treese, Dan Wallach, Avishai Wool, Karen Yannetta, and Michal Zalewski, among many others. BILL CHESWICK STEVE BELLOVIN AVI RUBIN 020163466XP01302003

Firewalls and Internet Security

Abstract: In a process for the manufacture of 2-amino-3-bromoanthraquinone by heating 2-aminoanthraquinone with bromine (in the molar ratio of 1:1) in sulfuric acid, while mixing, the improvement wherein crude 2-aminoanthraquinone, in sulfuric acid of from 60 to 90 percent strength by weight, which contains from 10 to 15% by weight of an alkanecarboxylic acid of 3 or 4 carbon atoms or a mixture of such acids, is heated with from 1 to 1.05 moles of bromine per mole of 2-aminoanthraquinone at from 130 to 150 DEG C. The 2-amino-3-bromoanthraquinone which is isolated may be used for the manufacture of dyes. It is at least as pure as that obtained from purified 2-aminoanthraquinone by the process of the prior art.

Network firewalls

Abstract: Computer security is a hard problem. Security on networked computers is much harder. Firewalls (barriers between two networks), when used properly, can provide a significant increase in computer security. The authors classify firewalls into three main categories: packet filtering, circuit gateways, and application gateways. Commonly, more than one of these is used at the same time. Their examples and discussion relate to UNIX systems and programs. The majority of multiuser machines on the Internet run some version of the UNIX operating system. Most application-level gateways are implemented in UNIX. This is not to say that other operating systems are more secure; however, there are fewer of them on the Internet, and they are less popular as targets for that reason. But the principles and philosophy apply to network gateways built on other operating systems as well. Their focus is on the TCP/IP protocol suite, especially as used on the Internet. >

Privacy-Enhanced Searches Using Encrypted Bloom Filters

Abstract: It is often necessary for two or more or more parties that do not fully trust each other to share data selectively. For example, one intelligence agency might be willing to turn over certain documents to another such agency, but only if the second agency requests the specific documents. The problem, of course, is finding out that such documents exist when access to the database is restricted. We propose a search scheme based on Bloom filters and group ciphers such as Pohlig-Hellman encryption. A semi-trusted third party can transform one party’s search queries to a form suitable for querying the other party’s database, in such a way that neither the third party nor the database owner can see the original query. Furthermore, the encryption keys used to construct the Bloom filters are not shared with this third party. Multiple providers and queriers are supported; provision can be made for third-party “warrant servers”, as well as “censorship sets” that limit the data to be shared.

Privacy-enhanced searches using encryption

Abstract: Encryption with keys that form an Abelian group are used in combination with a semi-trusted party that converts queries that are encrypted with the key of a querier to queries that are encrypted with the key of the encrypted database, without knowing the actual keys. In an illustrative embodiment, encryption is done with Bloom filters that employ Pohlig-Hellman encryption. Since the querier's key is not divulged, neither the semi-trusted party nor the publisher of the database can see the original queries. Provision can be made for fourth party “warrant servers”, as well as “censorship sets” that limit the data to be shared.

Crowds: anonymity for Web transactions

Abstract: In this paper we introduce a system called Crowds for protecting users' anonymity on the world-wide-web. Crowds, named for the notion of “blending into a crowd,” operates by grouping users into a large and geographically diverse group (crowd) that collectively issues requests on behalf of its members. Web servers are unable to learn the true source of a request because it is equally likely to have originated from any member of the crowd, and even collaborating crowd members cannot distinguish the originator of a request from a member who is merely forwarding the request on behalf of another. We describe the design, implementation, security, performance, and scalability of our system. Our security analysis introduces degrees of anonymity as an important tool for describing and proving anonymity properties.

How to model an internetwork

Abstract: Graphs are commonly used to model the structure of internetworks, for the study of problems ranging from routing to resource reservation. A variety of graph models are found in the literature, including regular topologies such as rings or stars, "well-known" topologies such as the original ARPAnet, and randomly generated topologies. Less common is any discussion of how closely these models correlate with real network topologies. We consider the problem of efficiently generating graph models that accurately reflect the topological properties of real internetworks. We compare the properties of graphs generated using various methods with those of real internets. We also propose efficient methods for generating topologies with particular properties, including a transit-stub model that correlates well with the internet structure. Improved models for the internetwork structure have the potential to impact the significance of simulation studies of internetworking solutions, providing a basis for the validity of the conclusions.

Searchable symmetric encryption: improved definitions and efficient constructions

Abstract: Searchable symmetric encryption (SSE) allows a party to outsource the storage of its data to another party (a server) in a private manner, while maintaining the ability to selectively search over it. This problem has been the focus of active research in recent years. In this paper we show two solutions to SSE that simultaneously enjoy the following properties: Both solutions are more efficient than all previous constant-round schemes. In particular, the work performed by the server per returned document is constant as opposed to linear in the size of the data. Both solutions enjoy stronger security guarantees than previous constant-round schemes. In fact, we point out subtle but serious problems with previous notions of security for SSE, and show how to design constructions which avoid these pitfalls. Further, our second solution also achieves what we call adaptive SSE security, where queries to the server can be chosen adaptively (by the adversary) during the execution of the search; this notion is both important in practice and has not been previously considered.Surprisingly, despite being more secure and more efficient, our SSE schemes are remarkably simple. We consider the simplicity of both solutions as an important step towards the deployment of SSE technologies.As an additional contribution, we also consider multi-user SSE. All prior work on SSE studied the setting where only the owner of the data is capable of submitting search queries. We consider the natural extension where an arbitrary group of parties other than the owner can submit search queries. We formally define SSE in the multi-user setting, and present an efficient construction that achieves better performance than simply using access control mechanisms.

A security architecture for the Internet protocol

Abstract: In this paper we present the design, rationale, and implementation of a security architecture for protecting the secrecy and integrity of Internet traffic at the Internet Protocol (IP) layer. The design includes three components: (1) a security policy for determining when, where, and how security measures are to be applied; (2) a modular key management protocol, called MKMP, for establishing shared secrets between communicating parties and meta-information prescribed by the security policy; and (3) the IP Security Protocol, as it is being standardized by the Internet Engineering Task Force, for applying security measures using information provided through the key management protocol. Effectively, these three components together allow for the establishment of a secure channel between any two communicating systems over the Internet. This technology is a component of IBM's firewall product and is now being ported to other IBM computer platforms.

Searchable symmetric encryption: Improved definitions and efficient constructions

Abstract: Searchable symmetric encryption SSE allows a party to outsource the storage of his data to another party in a private manner, while maintaining the ability to selectively search over it. This problem has been the focus of active research and several security definitions and constructions have been proposed. In this paper we begin by reviewing existing notions of security and propose new and stronger security definitions. We then present two constructions that we show secure under our new definitions. Interestingly, in addition to satisfying stronger security guarantees, our constructions are more efficient than all previous constructions.Further, prior work on SSE only considered the setting where only the owner of the data is capable of submitting search queries. We consider the natural extension where an arbitrary group of parties other than the owner can submit search queries. We formally define SSE in this multi-user setting, and present an efficient construction.