Mario, My feedback is embedded below.
-- JG [cid:image001.png@01D84340.6739D180] James Gould Fellow Engineer jgo...@verisign.com<applewebdata://13890C55-AAE8-4BF3-A6CE-B4BA42740803/jgo...@verisign.com> 703-948-3271 12061 Bluemont Way Reston, VA 20190 Verisign.com<http://verisigninc.com/> From: Mario Loffredo <mario.loffr...@iit.cnr.it> Date: Monday, March 28, 2022 at 6:59 AM To: James Gould <jgo...@verisign.com>, "regext@ietf.org" <regext@ietf.org> Subject: [EXTERNAL] Re: [regext] Comments to the feedback about epp-over-http Hi James, thanks for ypur quick reply. Please find my comments below. Il 25/03/2022 16:45, Gould, James ha scritto: Mario, For #4 “Cookie vs. HTTP Connection”, you asked the question “can you further clarify why we should opt for establishing the cookie at setup of the connection and how should it be possible? For example, what kind of request should be used to start the HTTP connection?”. I implemented pluggable transports in the Verisign EPP SDK, which included HTTP, HTTPS, TCP, and TLS. The Verisign EPP SDK does include a client interface as well as a server stub implementation, so I was able to see the transports from both sides. Support for HTTP and HTTPS was removed once we stopped supporting EPP over HTTP. The cookie is setup at the time of the HTTP or HTTPS connection. There is no “request” that is used to start the HTTP connection, just like the case for TCP and TLS. A network connection is made, which includes the underlying TLS handshake in the case of HTTPS and TLS and the cookie is setup for HTTP and HTTPS, and then the EPP application protocol rides on top of it. It seems to me a very low-level implementation. Sorry but I stiil don't understand how cookies are returned. They should be returned in an HTTP response that doesn't correspond to an HTTP request. Correct? JG: HTTP is being defined as a transport for EPP, so it should be treated as such. The setup and tear-down of the HTTP connection, includes the setup of the cookies that is used for HTTP session tracking. When a connection is made to the EPP / HTTPS server, the TLS connection is first established via the TLS-handshake and the HTTP session is established via setup of the session cookie (e.g., JSESSIONID), which then enables the server to maintain state. State such as tracking failed login attempts and timeouts (idle / command / absolute), which will result in a drop in the HTTP session / connection when the login attempt threshold is exceeded. At the application layer, a HTTPS connection is treated in the same way as a TLS connection, where the transport handles the framing and the connection state, and the application layer returns the EPP Greeting when the connection is setup and drops the connection with a EPP logout occurs or other EPP session conditions (e.g., exceeding timeouts). The HTTP session cookie is returned in all the HTTP responses. There is absolutely no intermingling of the transport layer (HTTP) with the application layer (EPP) other than making a connection available to read and write framed packets and the ability to close the connection when needed. If so, it seems to me uncompliant with what is stated in the first paragraph of Section 2.1 of RFC7230: HTTP is a stateless request/response protocol that operates by exchanging messages (Section 3<https://secure-web.cisco.com/1HLyHB_XdFEjfXQwGFkG0LZSLLnG4WFNvmZ-TNjfY6OfIaLTmL1biILbx72S_U5GmIiJtbpg0x39TJ52od6XvJJ-XGkebzBiEggAwQ42v85iZAXdc75eWbMUFwdy__GSQvGqoc_3jBcXvzaJksIZ9kIaGCObuN4sNz3NQqB4Gsr44Gdw9c4U6nDx9j_3kyp9hk3Za8zxK-CiPfgUAyNLhZ53ksc5Ae9PHvSSpW9kjbB9txLir0tvwL_pUPzDpJS2Y/https%3A%2F%2Fwww.rfc-editor.org%2Frfc%2Frfc7230%23section-3>) across a reliable transport- or session-layer "connection" (Section 6<https://secure-web.cisco.com/1vvJc-KSBR_KweDtRuo-dlLx6E_Z9fnj9TRqKJPKmI-z_MzpIyDJWDzfwvS180O64nU4M_K-HwHtvVlPkTyfnMVQ8I50S_dOCSrwzrW9QXz8ckfa9Yj0UZmgPk4aOBSCgpizCtKhPiufait2c-bik8SwDi5FGriXh2QetIDOP_N5qvkVGUCrJ7Agu0Kz55-5IRTrJpPK2rtdg6AYpfgj-5oaNCOjZ9kZcI233HN_HumLXDVcKmYklOTTqh_7mGPwt/https%3A%2F%2Fwww.rfc-editor.org%2Frfc%2Frfc7230%23section-6>). An HTTP "client" is a program that establishes a connection to a server for the purpose of sending one or more HTTP requests. An HTTP "server" is a program that accepts connections in order to service HTTP requests by sending HTTP responses. JG: I don’t see how using HTTP as a EPP transport would be considered uncompliant with RFC7230. Below is the flow of using HTTP as an EPP transport: 1. Establish TCP connection 2. Establish TLS session via TLS-handshake 3. Establish HTTP session via setup of the HTTP session cookie (e.g., JSESSIONID) 4. Return EPP Greeting in framed HTTP response with the session cookie (e.g., JSESSIONID) 5. Support HTTP requests in the form of framed EPP commands that are returned in HTTP responses in the form of framed EPP responses. 6. EPP session ends that drops the connection (HTTP / TLS / TCP) Anyway, the approach normally followed by every HTTP implementer leverages the services provided by an application server that incorporates an HTTP server. You don't need to manage HTTP connections, HTTP resquest/response marhsalling/unmarshalling and other low-level details. You just need to focus on the core of your server. Similarly, on the other side side, after having configured an HTTP client instance, a client simply builds and sends an HTTP request, and then receives and processes an HTTP response. JG: An EPP client or server should be capable of reusing all the same command and response processing without dealing with transport specific logic. The transport layer can be built as a layer that is pluggable. A client can configure the transport to use on a per-server basis, and the server can reuse the same command processing logic whether running on top of TLS or HTTPS, since TLS and HTTPS are purely transports. If there is a HTTP transport for EPP, then it should be purely a transport and not intermingle with the application protocol. There is no need to directly tie a HTTP session with the established EPP session. By keeping them separate, it makes it much easier for the client and the server to have pluggable transports, where the processing of the EPP commands including the hello, login, and logout treat the transport as simply a connection with no intermingling of session management. If it means that you have to reimplement what has commonly been used for years, it isn't clear to me how it could be much easier. On the contrary, it seems to me much trickier, especially if you implement EPP-over-HTTP from scratch. If the goal is using the same approach for mapping every transport layer without considering the peculiarites of each transport protocol and the technology the implementers are accostumed to use to deal with them, it doesn't make sense to me. For example, I cannot figure out how HTTP load balancing could support your solution. JG: The HTTP transport for EPP can easily be defined as a pure transport, which provides a connection for the application layer to leverage in the same fashion as TLS. The load balancer can leverage the TLS session id to route to the EPP gateway and in the case of HTTP, the HTTP session can be used to route the request to the appropriate application server that contains the connection state. Some HTTP application server proxies leverage the HTTP session to route the requests to the application server that contains the HTTP session and state. You need to establish the HTTP session when the connection is established to support the needed routing. There are other load balancing options, but by simplifying EPP over HTTP to have HTTP as purely a transport, it provides maximum flexibility in routing and maximum pluggability of transports (TCP, TLS, HTTP, HTTPS). Best, Mario -- JG [cid:image002.png@01D84340.6739D180] James Gould Fellow Engineer jgo...@verisign.com<applewebdata://13890C55-AAE8-4BF3-A6CE-B4BA42740803/jgo...@verisign.com> 703-948-3271 12061 Bluemont Way Reston, VA 20190 Verisign.com<http://secure-web.cisco.com/1ojxKstUEceETR7pe4khQ2dcwgGUwoOuZu4vhcT-TZcTCCpM93jXl6eq5i46tfTPt6lE9M6FO7hPqFZQVgxN5su1ozpEIXrOmp6SzB5kEB5SXCcRaJIg-vRb2B_SxGeDcgjxTqmZNEVO_Z0LtwZEFTLbzsLhmMMT-MlZ9pX7rpBY6Q8K0RxXyklhUEXCOqlJ0RMcZX2tEZNHp0Pr6XFhgw4X7OA0XCs0tk_jqiJbYMrkAFF0PXiAItfCmm_MR8whC/http%3A%2F%2Fverisigninc.com%2F> From: regext <regext-boun...@ietf.org><mailto:regext-boun...@ietf.org> on behalf of Mario Loffredo <mario.loffr...@iit.cnr.it><mailto:mario.loffr...@iit.cnr.it> Date: Friday, March 25, 2022 at 11:01 AM To: "regext@ietf.org"<mailto:regext@ietf.org> <regext@ietf.org><mailto:regext@ietf.org> Subject: [EXTERNAL] [regext] Comments to the feedback about epp-over-http Hi folks, here are in the following some comments grouped by subject to last meeting's feedback about EPP-over-HTTP: 1) Draft title Ulrich suggested to call the document EPP-over-HTTPS. I replied that the name was assigned to be consistent with RFC5734, i.e. EPP-over-TCP. SImilarly to RFC5734, the draft states, first in the abstract and then in the security considerations, that TLS is required. That being said, the authors don't object to renaming the dcocument EPP-over-HTTPS if the WG agrees. 2) Cookies Jim (Reed) asked why cookies should be used in this case. The reasons why we have used session cookiea are that they represent a standard method (RFC6265), well known to the community of REST service implementers, largely used and natively supported by libraries and frameworks on both client and server side. For example, it is the same method used by rdap-openid to map an RDAP session and tie it to an access token :-) .it and .pl have been using this method since the beginning and the registrars, after being informed that they had to enable cookies in their HTTP clients, have no longer complained about cookie management. In addition, the implementation of such a method doesn't introduce any change to the EPP core spec. Indeed, it preserves EPP comands semantics and doesn't mix the application layer with the transport layer. I would like to say that, regarding the clear distinction between those layers, this proposal is even better than RFC5734 as every EPP response is returned by the server as a consequence of receiving an EPP request. On the contrary, in RFC5734, an EPP <greeting> is returned to the client after the TCP connection has been established so, at least in this case, the two layers get mixed. Which method other than session cookie shoud be used instead ? 3) Security Considerations Ulirch recommended to review the security considerations by inheriting those from TLS WG about which versions and ciphers of TLS to use. Thanks a lot for the heads up, Ulrich. Surely, we'll do. Gavin noted that, unlike EPP-over-TCP, this draft states that client IP address check is optional. As a matter of fact, it is stated as recommended. Anyway, the authors don't object to changing it into an absolute rquirement if the WG agrees. 4) Cookie vs. HTTP Connection One comment from James in the chat is about establishing the cookie at setup of the connection and not linking it to the EPP Login command. James, can you further clarify why we should opt for establishing the cookie at setup of the connection and how shoudl it be possible? For example, what kind of request should be used to start the HTTP connection? IMO, an HTTP session is something that is inherently unlinked to the HTTP connections. HTTP connections can be broken but sessions don't get lost. Programmatically, REST implementers are in charge of processing HTTP requests and building responses rather than managing HTTP connections, which is instead delegated to the application servers. Finally, I would like to outline that Section 2.9.1 of RFC5730 states that an EPP session starts with a Login command and the mechanism described by RFC6265 lets (I'm quoting here) "the servers maintain a stateful session over the mostly stateless HTTP protocol". As a consequence, it seems much more practical to start the EPP/HTTP session as a result of a Login command. 5) EPP/HTTP Sessions vs. HTTP3 Connections Ulrich remarked that, in HTTP3, it is possible to have multiple sessions on an HTTP connection. This is valid also for the other HTTP versions. In fact, an HTTP connection can be kept alive and, over it, a client could submit multiple login-commands-logout sequences. This is quite usual for a smart client managing a pool of HTTP connections. Instead, It is unlikely but not impossible to come across HTTP connections supporting multiple concurrent sessions. What should be the possible drawbacks for a server in allowing the scenarios above? 6) Client authentication in HTTP3 Another note pointed out that HTTP3 client authentication requirements are different from this draft and they need to be reconciled. Think that it could be sufficient to add to the security considerations some text similar to what is included in section 4.4 "Peer authentication" of RFC 9001 "Using TLS to secure QUIC": A client MUST authenticate the identity of the server. This typically involves verification that the identity of the server is included in a certificate and that the certificate is issued by a trusted entity (see for example [RFC2818<https://secure-web.cisco.com/1tkPqD6aAMfTp6t6KppHDGdn0xpgOWSMeCqwZR-hs9Duph9SiceA0vCbaJJVqywNzcZscc6vUZkCPH9lCdtXq_nVJ3OfnNMz-XTQvhBuSBLsXa0VfO5ZNrGo47fSin8GRaQOWRSvSQP_7vRecBdddhF6L0Yqx3KcxAGpvdxpFCsPhLcd5bm-aVy3vTko6TfGBlohe2XEw3bwUmH-u56ZuZz50CUXUqA3YDFgBENAwPrLBjwdDxwgG1JFVGPZF_LEQ/https%3A%2F%2Fdatatracker.ietf.org%2Fdoc%2Fhtml%2Frfc2818>]). The draft has only considered the optional use of a certificate on server side (not on client side). In doing that, the draft is consistent with another sentence in the same paragraph of RFC9001: A server MAY request that the client authenticate during the handshake. A server MAY refuse a connection if the client is unable to authenticate when requested. Would it address the feedback? That's all for now. Hope I did not miss anything. Thanks a lot for your interest and feedback. Looking forward to your further comments. Best, Mario -- Dr. Mario Loffredo Technological Unit “Digital Innovation” Institute of Informatics and Telematics (IIT) National Research Council (CNR) via G. Moruzzi 1, I-56124 PISA, Italy Phone: +39.0503153497 Web: http://www.iit.cnr.it/mario.loffredo<http://secure-web.cisco.com/1Af0lIvD4wj2TkngfEVwrFVEVRozLiVIU2m1Gi1AH1GN4-eWP-IREXsOLCh8OJ03YAxRNYqVCPHQSwgj-tMCyzvN3jokOBTIBx4-5n1pJdiRXxl-ShJCaHkCjGNPa8EA5qw4kPjkxIrFAy1qOTcPFhieqdc_xyu8yisYoXzily9ozVw3GZaUtkKrLnmnDJhlFv2LRTCTnw913LzH8bX-hB6FpPlyFi_0v2_H1NFCgZYjuu4pgUeOeIqQJRQwtzNLS/http%3A%2F%2Fwww.iit.cnr.it%2Fmario.loffredo> -- Dr. Mario Loffredo Technological Unit “Digital Innovation” Institute of Informatics and Telematics (IIT) National Research Council (CNR) via G. Moruzzi 1, I-56124 PISA, Italy Phone: +39.0503153497 Web: http://www.iit.cnr.it/mario.loffredo<http://secure-web.cisco.com/1OpUT5efcRPCKu9ZmpfyFmbeoIexTqMIlduFKl7VTUSeET1W3oogufCAfJPPw-WSQ0iyo_wkztY_7M69qw0yf13C4w7OWB_Zk5w2FsV-GGCSLot8lFQdslU0DvCpCTw9FjZr0AsFsubGH-kToYrLxzzEj7vGDKuxIFxPK92P5bu2CDOE4UFtRPRGvu9rzX83rks8X2le_j1uSS1SD5B8CyvSp0tsne0HWMS5luho_DIhdOil7-0a_7ZX6CQA08GQq/http%3A%2F%2Fwww.iit.cnr.it%2Fmario.loffredo>
_______________________________________________ regext mailing list regext@ietf.org https://www.ietf.org/mailman/listinfo/regext
