Re: [TLS] Industry Concerns about TLS 1.3
nalini.elk...@insidethestack.com writes: > [ Unknown encryption status ] > [ Unknown signature status ] > > > >>> >>> What I am saying, in relation to your "Delivering a stable product" >>> comment is that over time various industries have learned what it takes to >>> "Deliver a stable product". We did not >>want to invest millions in >>> these debugging networks. But we learned the hard way, that it was >>> necessary. >>> I am not a member of the banking coalition that started this subject, nor >>> of the banking industry at all, but I certainly understand their >>> perspective and am concerned about the same >>unmanageable future they >>> described. > >>Do Akami, Cloudlflare and Google magically not have these problems? > It would be very interesting to get the network diagnostic and > operations people (rather than the architects) of the above companies > involved in this conversation. Hi, technical person most directly responsible for incident response and urgent debugging here. We modify endpoints to get what we need. We did have taps that relied on knowing the RSA Kx secret... but haven't used them in about a decade. I think the banks have an answer not available to the global passive adversaries: modify the server or client to use a fixed ECDH share, then use tech much like their current choices. It'll take a while to develop, but nobody in that environment plans to move to TLS 1.3 for operational systems any time soon anyway. -Brian > Also, you know, companies don't really enjoy spending money on network > diagnostic products which might be considered overhead. So, if they are, we > might do them the courtesy of not thinking that they are foolish to do so. > Why don't we listen to each other? I know at IETF, I often hear that we > don't get enough operators to comment and give feedback. Well, here you have > some. It may be that these companies have problems that are different from > Google's (just as an example). > Isn't our goal to have the best standards possible? Any organism (including > the IETF), needs feedback to thrive. > Nalini >> >> Thanks >> >> Mike >> >> >> >> -Original Message- >> From: Jeffrey Walton [mailto:noloa...@gmail.com] >> Sent: Friday, September 23, 2016 10:55 AM >> To: Ackermann, Michael>> Cc: BITS Security ; tls@ietf.org >> Subject: Re: [TLS] Industry Concerns about TLS 1.3 >> >> On Fri, Sep 23, 2016 at 10:46 AM, Ackermann, Michael >> wrote: >>> From the perspective an Enterprise that runs these applications and has >>> invested HEAVILY in the debugging networks. >>> >>> The reason we are debugging these networks is so that "The 5-6 order of >>> magnitude of folks using them" will have good service. If they do not, >>> they will consider competitors and/or generate a litany service calls or >>> complaints. I.E. When these "Folks" are slow or not working they >>> are just as unhappy as we are. >>> >> >> Isn't that the market operating as expected? Those who deliver a stable >> product at a competitive price are rewarded, while those who fail to deliver >> or deliver at an unreasonable cost are not? (Some hand waiving). >> >> If all providers failed to deliver or delivered an inferior product, then it >> might indicate a major course correction is needed. But I don't think that's >> the case here. >> >> Jeff >> >> >> The information contained in this communication is highly confidential and >> is intended solely for the use of the individual(s) to whom this >> communication is directed. If you are not the intended recipient, you are >> hereby notified that any viewing, copying, disclosure or distribution of >> this information is prohibited. Please notify the sender, by electronic mail >> or telephone, of any unintended receipt and delete the original message >> without making any copies. >> >> Blue Cross Blue Shield of Michigan and Blue Care Network of Michigan are >>nonprofit corporations and independent licensees of the Blue Cross and Blue >>Shield Association. >> ___ >> TLS mailing list >> TLS@ietf.org >> https://www.ietf.org/mailman/listinfo/tls > > > > -- > "Man is born free, but everywhere he is in chains". > --Rousseau. > > ___ > TLS mailing list > TLS@ietf.org > https://www.ietf.org/mailman/listinfo/tls > > > > ___ > TLS mailing list > TLS@ietf.org > https://www.ietf.org/mailman/listinfo/tls > > > >>> >>> What I am saying, in relation to your "Delivering a stable product" >>> comment is that over time various industries have learned what it takes to >>> "Deliver a stable product". We did not >>want to invest millions in >>> these debugging networks. But we learned the hard way, that it was >>> necessary. >>> I am not a member of the banking coalition
Re: [TLS] Industry Concerns about TLS 1.3
On Thu, Sep 22, 2016 at 03:29:42PM -0400, Dave Garrett wrote: > > Yes, all of these other channels are protected using TLS... which you > do not control in any way. Also, many sites/services already prioritize > FS cipher suites, so the deprecation of plain RSA key exchange doesn't > actually affect the vast majority of people. (e.g. Facebook & Twitter > both prefer ECDHE with NIST P-256) Within this very argument is > already the argument that supervision at endpoints is required here. > he security on the pipe is irrelevant. I don't see how you can make a > point to bring this up but think keeping plain RSA KE suites is a > useful solution. Well, if you have client and server that both have RSA KE enabled (even if not preferred), you can do something like this: - Intercept ClientHello. Strip any FPS ciphersuites and any associated extensions, strip EMS and ALPN but keep client_version and client_random the same. - The server will select RSA KE. - Intercept Certificate message and replace it with MITM cert. - Intercept ClientKeyExchange decrypt the key and re-encrypt with the original server key. - Compute the session key and log it. - Intercept the ClientFinished message and replace it with fixed version. - Intercept the ServerFinished message and replace it with fixed version. - Get off the active path, becoming passive. - Log the data transfer on the connection, and later decrypt it if desired. The keys will match, because they only depend on randoms (which were preserved) and the encrypted secret (which was preserved too). The EMS was stripped because it would break this. The MITM proxy knows the secrets needed to fix the Finished messages. Preferring FS won't help, because MITM proxy can downgrade the crypto negotiation. (And it is not like one can easily get the keys without either replacing the cert or knowing the RSA private key corresponding to it). -Ilari ___ TLS mailing list TLS@ietf.org https://www.ietf.org/mailman/listinfo/tls
Re: [TLS] Industry Concerns about TLS 1.3
What I mean is that we have Many MITM solutions today and they are able to be a good source for troubleshooting/diagnostics, in limited situations or perspectives.This lack of scope, depth and detail are what drove us to install the packet collection infrastructures (debugging networks I think some are saying). Some of the issues we have found include: * Level of detail is not sufficient. * Inherent tools are not sophisticated. * Data is difficult to share. * MITM management systems frequently do not "Play well" with overall management frameworks and other tools. * Problems take longer to resolve. * Depending upon the level of information logged on an MITM platform, the inherent processing can (and frequently does), create performance problems on the MITM and can change conditions, obfuscating the original issue. * MITMs are not usually configured to record, retain, archive or manage large amounts of diagnostic data. * A MITM platform frequently has a very limited vision of the overall session path. Or, there may be multiple MITM's involved in a session path.Which one has the best view? Which one(s) should you focus on? * AND, as you said, the more MITMs you add, the more latency and complexity you are forced to deal with.As a network diagnostic person this approach can actually make your responsibilities more difficult and less achievable. * The information collected is usually not granular enough to perform network diagnostics for those situations that require it. * One key initial piece of information in troubleshooting is to determine if the source of the problem is in: the Client, the Network or the Server. MITM's are rarely able to provide insight into this critical and time sensitive determination. * in general diagnostics are made more difficult with this approach. Multiple sessions and possibly interfaces may need to be traced and the MITM can further confuse things by acting as a router and/or a nat and/or a Load Balancer. As someone who deals with these types of additional complexities every day, I would like to see fewer MITMs rather than more. My organization uses the MITM diagnostics and management systems, and like most point solutions they are a valuable facet of our diagnostic arsenal, but because of the manifold shortcomings (some of which are listed above), they are not a central focus and are not a viable initial focal point or suitable overall point of triage. Triage may dictate the use of the MITM diagnostics, but MITM troubleshooting/diagnostics would not be an effective method on its own, for most situations/issues that exist in today's complex multi-tier applications. My job is to make things work and fix them ASAP when they don't. While I fully understand the need for effective security, I hope those on the security side would conversely understand the need to make things work, perform well, and quickly diagnose problems when they do not. I am often reminded of a colleague who states (jokingly I think) "The ultimate security is where no data flows and nothing works". I hope this is not the direction we are heading and that some form of compromise will be forthcoming between these two discrete factions with differing perspectives. -Original Message- From: TLS [mailto:tls-boun...@ietf.org] On Behalf Of Pawel Jakub Dawidek Sent: Saturday, September 24, 2016 2:54 AM To: tls@ietf.org Subject: Re: [TLS] Industry Concerns about TLS 1.3 Hello. As a vendor of one of those MITM proxy solutions for TLS inspection I'd be grateful if you could be more specific about the "scope, depth and detail" that is not delivered by such solutions, so we can have discussion whether this is something we can address or not and if not, maybe we can come up with some alternative solutions before we give up. By doing MITM we increase latency, even if very little, that's inevitable. But can you really avoid doing MITM TLS inspection? In my opinion, no. Let me elaborate. Of course the amount of TLS traffic is growing rapidly (which is a good thing) thanks to many "contributions": - Server Name Indication extenstion, - Edward Snowden, - Free certificates (eg. Let's Encrypt), - HTTP 2.0. Because of that, every corporate network needs visibility inside TLS traffic not only incoming, but also outgoing, so they can not only debug, but also look for data leaks, malware, etc. Customers are increasingly aware of all this and it is not a question of MITM incresing latency, because it has to be done, the more important is to make it in a decrypt-once-feed-many fashion, as they have multiple solutions in place to analyze the traffic for different reasons. And when you do data leak prevention or malware detection you want to be in the middle so you can terminate the session as quickly as possible. I don't want to say that Forward Secrecy comes at no price. It makes
Re: [TLS] Proposed Change to Certificate message (#654)
On Sat, Sep 24, 2016 at 09:31:51PM +1000, Martin Thomson wrote: > On 24 September 2016 at 19:17, Ilari Liusvaara> wrote: > > It occured to me that certain extensions might be considered to be per- > > chain. Like e.g. type of the certificate. Where do extensions like that > > go? Always to the extension block of the first certificate (except that > > might cause somewhat of a cyclic dependency in parsing)? > > The type of which certificate? The end-entity? Seems like that > belongs with the end-entity cert then. I mean equivalent of the client_certificate_type/server_certificate_type extensions. And the way those extensions are defined, those scope the entiere chain. E.g. There was some discussion about "subcerts"[1]. One way to add those would be as a new certificate type. ... Or are new certificate types like new CLASSes in DNS: Heavy objects dropped by bad idea fairy? :-> But in the future, there might very well be new extensions that are scoped to certificate chain and not and individual certificate. And those can't be put into EncryptedExtensions if server can send multiple certificates (like it can in post-handshake auth extension) or if client needs to send one. [1] The usecases can't be practically accomplished today: The mode of operation is just plain alien to X.509. -Ilari ___ TLS mailing list TLS@ietf.org https://www.ietf.org/mailman/listinfo/tls
Re: [TLS] Proposed Change to Certificate message (#654)
On Fri, Sep 23, 2016 at 11:05:10PM +, Nick Sullivan wrote: > Thanks for the suggestions. I've restructured my PR to include an array of > SingleCertificate objects in the Certificate structure. It occured to me that certain extensions might be considered to be per- chain. Like e.g. type of the certificate. Where do extensions like that go? Always to the extension block of the first certificate (except that might cause somewhat of a cyclic dependency in parsing)? And then there is the user_mapping. I presume mechanism like this is to be used to transport it (avoiding need to mess with new handshake messages and such. > Ilari: I agree that the post-hanshake auth mechanism as currently described > is a bit lacking, but I'd like to sort this out first. Well, more like I was annoyed at having to implement that at all and the fact that it requires remembering a hash state (which may be a quite harsh requirement in some cases). -Ilari ___ TLS mailing list TLS@ietf.org https://www.ietf.org/mailman/listinfo/tls