Older: 
Security built in, not bolted on
Location: Flying over Atlantic Ocean, Amsterdam to Boston
Am flying back from a week’s vacation in Europe with family and friends. While sorting through a huge number of emails I have neglected (or ignored) while on vacation, I ran across a note from the leader of my Security research and architecture team at Nortel. He forwarded me an article from CNET that references a recent talk by industry security guru Bruce Schneier, who laments the very existence of the security industry, pointing out that our need to layer security products on top of various core transport, compute and productivity devices and applications is a clear indictment that those very entities are insecure and incomplete.
Reading this article struck me as a bit of déjà vu because prior to joining Nortel and Broadcom, I spent over 10 years of my career developing systems and technology that brought security into the very essence of communications transport. We, as an industry, tend to forget that the best networks and applications provide their core services in a way that fosters a high level of use but that also acknowledges that with that use comes inevitable abuse by people who are either malicious or just simply opportunistic.
If we think about the communications systems of the IBM SNA or Digital VAX era, what we had were very mature, complete systems that were able to deliver technology in context to the risks of their small controlled world. When the industry transformed to the Ethernet and PC era in the early 1990’s, we forgot about the kind of control the mainframe era had and, as a result, lost most elements of security and predictability. We did not particularly care, however, because there were no real defined threats at that time.
But, by the late 1990’s, into the early 2000’s, we found that we had become so dependent on these fantastic networks and internetworks, which were the source of huge productivity gains, that when the inevitable happened, and mass attacks on that infrastructure emerged, we were stunned and forced to evolve. With CodeRed, Nimda, the Love Bug and a host of other Internet and email worms and viruses, we realized that the nice simple, high-performance, global Internet was so unintelligent that it could barely detect, much less defend, itself from these attacks. From 2000 through to today we refocused, as an industry, on security in any form that would protect us.
So, here we are at the edge of a new era, one driven by hyperconnectivity. An era that will provide connectivity to a more diverse set of devices than ever before, one that will ingrain communications into every aspect of our lives and the systems we interact with.
The challenge that we face, though, is that, much as Bruce points out, the bolted-on security of our systems is a drag on their effectiveness and a mask over their true issue - the lack of strong inherent security capability. The technologies and systems that make up the next era of communications and IT must not forget the lesson of the past: that great networks must defend themselves because they will inevitably be the target of misuse, simply because they are great networks.
Looking at some of the technologies that we see as key to this future and their security posture, there seems to be some hope that we are starting from a stronger position. It is entirely possible, however, that the industry, much as it has done before, will be swept up in the rush to connect everything and to drive down cost that we will forget that the integrity of that system is an inherent requirement that cannot be compromised, ever.
As an example, in the 4G wireless dialog, there are a few technologies in the radio access network space. These include IEEE 802.16e Mobile WiMAX, 3GPP LTE, and 3GPP2 UMB (or CDMA RevC). Most of the debate in the industry is centered on which of these radio networks is the “right choice,” and far too little dialog is being focused on the other characteristics of the technologies. For those of you familiar with WiMAX or LTE, or even GSM or UMTS, ask yourselves the following questions. Do you understand the security model of each? Do you understand if it is strong or weak? Do you understand what assumptions were made about the system and how it will behave from a security perspective? Do you understand the cost of the security model used and its impact as overhead?
I don’t ask these questions for any other reason than to point out that it is easy to get caught up in transport technology or “speeds and feeds” and ignore the other aspects of a well-rounded communications system, which would result in an inevitable pattern of repeating our past mistakes. One of the reasons I feel that WiMAX 802.16e (not just the RF part but the whole system) has a strong possibility of playing a major role in the 4G hyperconnected future is that in addition to being a high-performance, low-cost broadband transport system, it leverages the IEEE security learnings (specifically the trials and errors of the Wi-Fi world) to deliver a system that makes few assumptions about the end systems’ security, but realizes that it must control access, protect transport and remain resilient even in the presence of the unknown.
We are also trying to make sure that the other 4G technologies, like LTE, take into account the hyper-connected world and its inevitable diversity and scale, and that security is architected in from the beginning. We shall see if we are smarter as an industry this time around and if we ultimately build the systems of the next generation to not only connect things but to also protect them with embedded security capability. This will simplify the system by eliminating the bolted-on security of past systems and maybe even achieve the end state that Bruce describes (or get us closer to it).
Let’s hope that one of the characteristics of 4G is that, from its inception, it’s a “Secure Network”.
[…] While I have a number of topics in the queue, I wanted to take a run today at a few of the comments and questions that emerged from my Security post. […]
May 24th, 2007 at 2:04 pm from John Roese’s Blog » Blog Archive » Some Q&A Catch Up