To: Frank A. Coluccio who wrote (2457 ) 12/6/1998 3:20:00 PM From: ftth Read Replies (3) | Respond to of 12823
Hi Frank, I second the thanks for all your contributions to this thread. There are several threads that discuss future communications technologies, but none as successful and enlightening as this one (in no small part to your posts). Before I found this thread, I tried to start one myself called Broadband Communications Technologies, but it died a quick death. As for security-related stuff, I'm not sure where to start because it encompasses many different topics. There are a host of entire books devoted to portions of the topic. In general, the average user considers security and encryption as synonymous, but a robust security solution is much more involved. I'm sure you already know most of this, but here's some basic internet security information that may provide a seed for comments by others. You'll see different names or groupings of the categories I've listed if you read security literature. Generally they only show 4-—maybe 5 categories, but I think these 7 cover all the major concerns. This topic is far too complex to accurately summarize without some generalities, so bear with me in that regard. These are the major categories which must be considered for a complete security solution (where all interested parties feel secure): 1. AUTHENTICATION—This provides assurance that the parties involved “are who they say they are” (i.e. to prevent fraud). Public key cryptosystems (actually, there's only 1 in widespread use today (RSA)) are one [part of a] solution to this category. 2. AUTHORIZATION—This provides assurance that the party has the proper access privilege for the product or service being requested; or, it is the process of assigning a “billing level” for the current session where a user may require different levels of service for different sessions (but they are authorized for all or most levels). Sort of a "dynamic authorization." 3. PRIVACY—This provides assurance that the product or service transaction cannot be observed (or hijacked and observed later) by non-authorized, non-authenticated parties (i.e. this is the encryption part of security). Note that simply possessing a crypto key should NOT serve as a substitute for the AUTHORIZATION and AUTHENTICATION steps. 4. DATA INTEGRITY—This provides assurance that the product or service was not modified/altered/substituted while in-transit (sort of a “digital tamper-proof seal”). 5. NONREPUDIATION—This provides assurances that the parties involved cannot deny they initiated a communication request/reply for a product or service (e.g. “I never ordered that” or “I never got your invoice”). 6. AUDITING—This provides the network security administrator with assurances that the security policies are working as desired, and that attempted breaches of security either aren't happening, or were stopped by a safety net. 7. TRUSTED THIRD PARTIES—This provides a certification agency (or Certificate Authority) that is known and trusted by all parties, and can be used to certify the validity of the parties involved (Digital Certificates), as well as providing an independent transaction processing center so that you don't have to worry about ordering from a “shell company” that just wants your credit card information. These 2 functions may or may not be through the same third-party. <<There are literally dozens of small companies (many not public) that have products to deal with some of these categories, as well as things that are further down the chain. As I've said before, bandwidth and security are "joined at the hip," so I'm quite convinced this category of companies is vital to the broadband future, and vise-versa.>> For some secure applications, coverage of all aspects isn't necessary at present for it to be considered “secure.” But, the “at present” part is important. If there are holes that can be exploited for profit, they will be. Just because a particular type of crime doesn't exist today doesn't mean it won't be invented. The criminals here are a bit more educated and clever that your average street thug. Look at the piracy history of European Satellite systems-—all the systems that would ‘never be hacked' have been. Primarily becasue they were only partial solutions or they didn't follow their own rules for auditing and key management. Another key point is that the 2 ends of a given transaction have different “needs” from a security standpoint. For example, the user/consumer could probably care less if the security model has coverage for anti-piracy (stealing the service provider's service (although in the end, this could cause their rates to increase—or at least serve as an excuse to that end)). Digital Signatures provide coverage for the Authentication, Data Integrity, and Non-repudiation portion. Message Digest functions and 1-way Hash functions are terms you'll see associated with Digital Signatures. A public key cryptosystem is also a necessary part of Digital Signatures. As a separate step, Digital Signatures can be applied to the Authorization steps to insure the validity of the authorization process. Whether this is used depends on the value of what is being authorized. For example, once all bandwidth becomes metered (it will), a clever user could set their connection for more that they are authorized for, while also modifying the audit trail so the audit doesn't show it. Digital Signatures can prevent this. As well, as mentioned in the example above, Digital signatures can be applied to the Auditing step to insure its integrity. False or modified values, or redirected security queries are possible without adding this on top of the basic “Auditing in the clear” step. Beyond the technical issues, there are also a host of legal issues, political issues, and business issues. We'll save those for another day. As for DOCSIS, the DOCSIS security architecture only deals with privacy in the Baseline Privacy Spec, and adds piracy protection to the “Full Security” (ha ha) spec. The DOCSIS Security Assessment Working Group actually recommended that the “Full Security” option be dropped, and that they merely add cable modem authentication to the Baseline Privacy Spec to give them sufficient security. This is just amazing to me, especially when you consider there were representatives from 17 different companies in this working group. I'm sure there are politics and economic considerations that drive this because such a basic system isn't technically sound (many potential holes). I guess it takes widespread losses and lawsuits before it's worth doing right. As for DAVIC, security wasn't added until DAVIC 1.3 (there might be a basic set of requirements in 1.2, I forget). However, it is a more complete solution and probably about as much as you could expect a group to agree on until such point that they begin to see the holes opening wider and the losses mounting. This is too long already. Nobody reads these ultra-long posts. dh