ALTO S. Kiesel Internet-Draft University of Stuttgart Intended status: Informational M. Stiemerling Expires: January 5, 2015 H-DA July 4, 2014 Application Layer Traffic Optimization (ALTO) Cross-Domain Server Discovery draft-kiesel-alto-xdom-disc-00 Abstract The goal of Application-Layer Traffic Optimization (ALTO) is to provide guidance to applications that have to select one or several hosts from a set of candidates capable of providing a desired resource. ALTO is realized by a client-server protocol. Before an ALTO client can ask for guidance it needs to discover one or more ALTO servers that can provide suitable guidance. In some deployment scenarios, in particular if the information about the network topology is partitioned and distributed over several ALTO servers, an ALTO client may need to discover an ALTO server outside of its own network domain, in order to get appropriate guidance. This document details applicable scenarios, itemizes requirements, and analyzes existing solution approaches for such ALTO cross-domain server discovery. However, the specification of a procedure is beyond the scope of this document. Note, that in earlier versions of this document, ALTO cross-domain server discovery was referred to as "third-party discovery", but it has been renamed to avoid naming ambiguities. Kiesel & Stiemerling Expires January 5, 2015 [Page 1] Internet-Draft ALTO Cross-Domain Server Discovery July 2014 Terminology and Requirements Language This document makes use of the ALTO terminology defined in RFC 5693 [RFC5693]. The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT", "SHOULD", "SHOULD NOT", "RECOMMENDED", "MAY", and "OPTIONAL" in this document are to be interpreted as described in RFC 2119 [RFC2119]. Status of this Memo This Internet-Draft is submitted in full conformance with the provisions of BCP 78 and BCP 79. Internet-Drafts are working documents of the Internet Engineering Task Force (IETF). Note that other groups may also distribute working documents as Internet-Drafts. The list of current Internet- Drafts is at http://datatracker.ietf.org/drafts/current/. Internet-Drafts are draft documents valid for a maximum of six months and may be updated, replaced, or obsoleted by other documents at any time. It is inappropriate to use Internet-Drafts as reference material or to cite them other than as "work in progress." This Internet-Draft will expire on January 5, 2015. Copyright Notice Copyright (c) 2014 IETF Trust and the persons identified as the document authors. All rights reserved. This document is subject to BCP 78 and the IETF Trust's Legal Provisions Relating to IETF Documents (http://trustee.ietf.org/license-info) in effect on the date of publication of this document. Please review these documents carefully, as they describe your rights and restrictions with respect to this document. Code Components extracted from this document must include Simplified BSD License text as described in Section 4.e of the Trust Legal Provisions and are provided without warranty as described in the Simplified BSD License. Kiesel & Stiemerling Expires January 5, 2015 [Page 2] Internet-Draft ALTO Cross-Domain Server Discovery July 2014 Table of Contents 1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . 4 1.1. Document History . . . . . . . . . . . . . . . . . . . . . 4 1.2. Feedback . . . . . . . . . . . . . . . . . . . . . . . . . 4 2. The Need for ALTO Cross-Domain Server Discovery . . . . . . . 5 2.1. Partitioned ALTO Knowledge . . . . . . . . . . . . . . . . 5 2.2. ALTO Queries on behalf of a Third Party . . . . . . . . . 6 2.3. Partitioned Knowledge and Queries for a a Third Party . . 7 2.4. Example . . . . . . . . . . . . . . . . . . . . . . . . . 7 3. Requirements for ALTO Cross-Domain Server Discovery . . . . . 9 3.1. Discovery Client Application Programming Interface . . . . 9 3.2. Data Storage and Authority Requirements . . . . . . . . . 9 3.3. Cross-Domain Operations Requirements . . . . . . . . . . . 9 3.4. Protocol Requirements . . . . . . . . . . . . . . . . . . 10 3.5. Further Requirements . . . . . . . . . . . . . . . . . . . 10 4. Related IETF Protocols and Activities . . . . . . . . . . . . 11 5. Security Considerations . . . . . . . . . . . . . . . . . . . 12 6. IANA Considerations . . . . . . . . . . . . . . . . . . . . . 13 7. References . . . . . . . . . . . . . . . . . . . . . . . . . . 14 7.1. Normative References . . . . . . . . . . . . . . . . . . . 14 7.2. Informative References . . . . . . . . . . . . . . . . . . 14 Appendix A. ALTO and Tracker-based Peer-to-Peer Applications . . 16 Appendix B. Contributors List and Acknowledgments . . . . . . . . 21 Authors' Addresses . . . . . . . . . . . . . . . . . . . . . . . . 22 Kiesel & Stiemerling Expires January 5, 2015 [Page 3] Internet-Draft ALTO Cross-Domain Server Discovery July 2014 1. Introduction Application-Layer Traffic Optimization (ALTO) [RFC5693] is realized by an HTTP-based client-server protocol [RFC7285]. Before an ALTO client can issue ALTO queries, it needs to discover a suitable ALTO server. During the design phase of the overall ALTO solution, two different server discovery scenarios have been identified and documented in the ALTO requirements document [RFC6708]. In the first scenario, documented in Req. AR-32, it is sufficient for a given ALTO client to discover a single ALTO server (or a small number of ALTO servers with identical knowledge, for reasons of reliability), usually the nearest one, as recommended by the operator of the access network. A discovery mechanism for this scenario, based on DHCP and DNS, is specified in [RFC7286]. An alternative approach, based on IP anycast, is documented in [I-D.kiesel-alto-ip-based-srv-disc]. In the second scenario, documented in Req. AR-33, an ALTO client may need to discover many different ALTO servers, depending on the queries it wants to issue. These ALTO servers may be located in other network domains than the client is. This document details applicable scenarios, itemizes requirements, and analyzes existing solution approaches for such ALTO cross-domain server discovery. However, the specification of a procedure is beyond the scope of this document. An experimental discovery procedure, which fulfills the requirements documented here, but which is currently not recommended for usage in the Internet, is documented in draft-kiesel-alto-xdom-disc-alg-00.txt. 1.1. Document History This document is a direct successor of [I-D.kiesel-alto-3pdisc] and [I-D.kist-alto-3pdisc]. The scenario and mechanisms described here and in these documents have been referred to as "third-party server discovery" in the past. However, to avoid ambiguities with a completely different scenario, it has been renamed to "ALTO Cross- Domain Server Discovery". 1.2. Feedback Comments and discussions about this document should be directed to the ALTO working group: alto@ietf.org. Kiesel & Stiemerling Expires January 5, 2015 [Page 4] Internet-Draft ALTO Cross-Domain Server Discovery July 2014 2. The Need for ALTO Cross-Domain Server Discovery ALTO Cross-Domain Server discovery is needed, if two independent effects appear at the conjunction: partitioned knowledge and clients sending queries on behalf of other hosts. The following two subsections will discuss these effects separately, before the next subsection will discuss the combination of both. 2.1. Partitioned ALTO Knowledge ALTO is realized by an HTTP-based client-server protocol. Basically, this protocol allows ALTO clients to download "cost maps", which indicate costs or other properties (according to various metrics) of the data path between endpoints. Furthermore, ALTO clients may access only parts of a cost map through the "map filtering service", and they may query endpoint properties through the "endpoint property service". Further similar services exist; for details refer to [RFC7285]. The endpoints are usually identified through their IP addresses. For efficiency reasons, endpoints may be grouped by using IP address prefixes and the "network map" with provider-defined identifiers (PIDs) defined within the ALTO protocol. Conceptually, the ALTO protocol allows to query for path costs between arbitrary IP addresses from the whole IP address range, i.e., a full 2^32 x 2^32 matrix for IPv4 and a 2^128 x 2^128 matrix for IPv6 (The "PID" mechanism introduced in the ALTO protocol makes storage and transmission more efficient but does not alter that basic principle). And in fact, there may be deployment scenarios where a single ALTO server (or a cluster of servers operated by a single organization) has this "Internet-wide view", e.g., a community project collecting end-to-end measurements. However, a very important class of scenarios is, when guiding information actually originates from the (access) network operators, such as Internet Service Providers (ISPs), IT departments of large companies or universities, etc. The information available at each of these providers will not be a full NxN matrix, but more like a 1xN vector, i.e., indicating cost "from us to anywhere" while cost "from anywhere to anywhere" is unknown. Several options exist how these pieces of information could be exposed to the ALTO clients, at least in theory: 1. The individual 1xN vectors could be aggregated to a virtual network-wide NxN matrix. This virtual matrix could be exposed through a single ALTO server or through a cluster of ALTO servers with identical information. However, no back-end protocol and process for this aggregation is currently defined. Kiesel & Stiemerling Expires January 5, 2015 [Page 5] Internet-Draft ALTO Cross-Domain Server Discovery July 2014 2. Each (access) network operator could operate their own ALTO server, which only has this partial knowledge available there. Requests for unknown (source,destination)-pairs would be redirected to another ALTO server. The idea has been discussed [I-D.kiesel-alto-alto4alto] but no complete specification exists. A related option would be to establish something similar to a web search engine, which would do the redirecting. But again, no detailed specification exists. 3. Each (access) network operator could operate their own ALTO server, which only has this partial knowledge available there. It would be the duty of the client-side ALTO server discovery mechanism to directly find an ALTO server that can reasonably answer a given query. This option will be considered in more detail in the remainder of this memo. 2.2. ALTO Queries on behalf of a Third Party An ALTO client may be co-located with or embedded into the resource consumer, i.e., the entity that seeks to access the desired resource and that will be one endpoint of the data transmission to be optimized. This kind of ALTO client will most often only be interested in "cost from me to somewhere else" queries, i.e., in an 1xN matrix. In contrast, the ALTO client may also be located at some kind of directory server, P2P tracker, CDN redirect server, etc., i.e., at some entity that takes part in the application signaling but is not the endpoint of the actual user data transmission to be optimized. From an ALTO perspective, this kind of ALTO client does not issue ALTO queries for its own optimization needs, but instead it issues queries on behalf of remote third parties. One motivation for this second type of configuration is faster deployment, as only some central servers would have to be equipped with an ALTO client and not all the clients. Furthermore, only these central servers (e.g., servers of a CDN) would need to access the ALTO servers, while the less-trusted clients could be denied to access the topology and cost maps. Another important reason is, that in some scenarios, much better optimization results can be achieved, if the ALTO guidance is considered at a central resource directory. See Appendix A for a detailed case study and analysis of such a scenario. In the second scenario, ALTO queries may be interested in the path costs from an arbitrary point in the network topology (where the third party is, on behalf of which the query is sent) to other arbitrary points in the topology (where the candidate resource Kiesel & Stiemerling Expires January 5, 2015 [Page 6] Internet-Draft ALTO Cross-Domain Server Discovery July 2014 providers are located). That is, such ALTO clients might want to access the full (virtual) NxN matrices. A more detailed discussion of various options where to place the functional entities comprising the overall ALTO architecture can be found in [I-D.ietf-alto-deployments]. 2.3. Partitioned Knowledge and Queries for a a Third Party As long as all ALTO servers store identical information, an ALTO client can send its queries to any server. A procedure for finding an ALTO server by means of DHCP is specified in [RFC7286]. If ALTO knowledge is partitioned and distributed over several ALTO servers and the ALTO client is co-located with a resource consumer, the DHCP based discovery procedure [RFC7286] will most likely work as well. The reason is, that this kind of ALTO client will only issue queries for "costs from me to anywhere", and the network operator can configure via DHCP an ALTO server that can answer these types of query. If, however, ALTO knowledge is partitioned and distributed over several ALTO servers and the ALTO client issues ALTO queries on behalf of third parties, a different kind of server discovery mechanism is needed. These ALTO queries will ask for the "costs from X to anywhere" (where X is the IP address or PID of the third party). For each of these queries a suitable ALTO server has to be found and X will be the parameter for the discovery mechanism. 2.4. Example The following, non-normative example illustrates the discovery procedure envisioned in this document. Assume a peer-to-peer tracker is located in the network operated by ISP A. Some peer, which is located in ISP B's network, asks the tracker for an ALTO-optimized list of other peers that take part in a specific swarm. The tracker can observe the source address X of this message, which is the peer's IP address. Assume that there is no omniscient ALTO server that knows the whole Internet topology. Therefore, the ALTO client in the tracker does a "back-connect" to the ALTO server operated by ISP B, which knows path costs from said peer (i.e., IP address X) to anywhere. The ALTO client retrieves this information and the tracker sorts the peer list according to it, before returning it to the peer. Before this "back-connect" can occur, the server ALTO discovery Kiesel & Stiemerling Expires January 5, 2015 [Page 7] Internet-Draft ALTO Cross-Domain Server Discovery July 2014 mechanism needs to map from the peer's IP address X to the ALTO server of ISP B, i.e., the network operator that controls IP address X and has assigned it to the peer. Kiesel & Stiemerling Expires January 5, 2015 [Page 8] Internet-Draft ALTO Cross-Domain Server Discovery July 2014 3. Requirements for ALTO Cross-Domain Server Discovery A solution for the problem described in the previous section would be an ALTO Cross-Domain Server Discovery system. This section itemizes requirements. 3.1. Discovery Client Application Programming Interface The discovery client would be called through some kind of application programming interface (API) and the parameters would be an IP address and, for purposes of extensibility, a service identifier such as "ALTO". It would return an URI that offers the requested service ("ALTO") for the given IP address. In other words, the client would be used to retrieve a mapping: (IP address, "ALTO") -> IRD-URI where IRD-URI is the URI of the Information Resource Directory (IRD, see Section 9 of [RFC7285]) of an ALTO server that can give reasonable guidance to a resource consumer with the indicated IP address. 3.2. Data Storage and Authority Requirements The information for mapping IP addresses and service parameters to URIs should be stored in a - preferably distributed - database. It must be possible to delegate administration of parts of this database. Usually, the mapping from a specific IP address to an URI is defined by the authority that has administrative control over this IP address, e.g., the ISP in residential access networks or the IT department in enterprise, university, or similar networks. 3.3. Cross-Domain Operations Requirements The cross-domain server discovery mechanism should be designed in such a way that it works across the public Internet and also in other IP-based networks. This in turn means that such mechanisms cannot rely on protocols that are not widely deployed across the Internet or protocols that require special handling within participating networks. An example is multicast, which is not generally available across the Internet. The ALTO Cross-Domain Server Discovery protocol must support gradual deployment without a network-wide flag day. If the mechanism needs some kind of well-known "rendezvous point", re-using an existing infrastructure (such as the DNS root servers or the WHOIS database) should be preferred over establishing a new one. Kiesel & Stiemerling Expires January 5, 2015 [Page 9] Internet-Draft ALTO Cross-Domain Server Discovery July 2014 3.4. Protocol Requirements The protocol must be able to operate across middleboxes, especially across NATs and firewalls. The protocol will specify an algorithm that determines the service parameters to be used when queries and responses are exchanged. This service parameter will specify 'ALTO' for the ALTO cross-domain service discovery. Potentially, it also specifies other required parameters needed for the service discovery, such as to be used transport or application level protocol. The protocol will support the query with the above mentioned service parameters and allow that the response contains one or more URI(s). The protocol shall not require any pre-knowledge from the client other than any information that is known to a regular IP host on the Internet. 3.5. Further Requirements The ALTO cross domain server discovery cannot assume that the server discovery client and the server discovery responding entity are under the same administrative control. Kiesel & Stiemerling Expires January 5, 2015 [Page 10] Internet-Draft ALTO Cross-Domain Server Discovery July 2014 4. Related IETF Protocols and Activities TBD. Survey. In particular, the ECRIT WG has specified a reverse- DNS-based procedure [RFC7216] to solve a similar problem; TBD: analyze whether we can re-use or adapt it. Kiesel & Stiemerling Expires January 5, 2015 [Page 11] Internet-Draft ALTO Cross-Domain Server Discovery July 2014 5. Security Considerations A high-level discussion of security issues related to ALTO is part of the ALTO problem statement [RFC5693]. A classification of unwanted information disclosure risks, as well as specific security-related requirements can be found in the ALTO requirements document [RFC6708]. The remainder of this section focuses on security threats and protection mechanisms for the third-party ALTO server discovery procedure as such. Once the ALTO server's URI has been discovered and the communication between the ALTO client and the ALTO server starts, the security threats and protection mechanisms discussed in the ALTO protocol specification [RFC7285] apply. TBD Kiesel & Stiemerling Expires January 5, 2015 [Page 12] Internet-Draft ALTO Cross-Domain Server Discovery July 2014 6. IANA Considerations This document does not require any IANA action. Kiesel & Stiemerling Expires January 5, 2015 [Page 13] Internet-Draft ALTO Cross-Domain Server Discovery July 2014 7. References 7.1. Normative References [RFC2119] Bradner, S., "Key words for use in RFCs to Indicate Requirement Levels", BCP 14, RFC 2119, March 1997. 7.2. Informative References [I-D.ietf-alto-deployments] Stiemerling, M., Kiesel, S., Previdi, S., and M. Scharf, "ALTO Deployment Considerations", draft-ietf-alto-deployments-09 (work in progress), February 2014. [I-D.kiesel-alto-3pdisc] Kiesel, S., Stiemerling, M., Schwan, N., Scharf, M., Tomsu, M., and H. Song, "ALTO Server Discovery Protocol", draft-kiesel-alto-3pdisc-05 (work in progress), March 2011. [I-D.kiesel-alto-alto4alto] Kiesel, S., "Using ALTO for ALTO server selection", draft-kiesel-alto-alto4alto-00 (work in progress), July 2010. [I-D.kiesel-alto-ip-based-srv-disc] Kiesel, S. and R. Penno, "Application-Layer Traffic Optimization (ALTO) Anycast Address", draft-kiesel-alto-ip-based-srv-disc-03 (work in progress), July 2014. [I-D.kist-alto-3pdisc] Kiesel, S., Krause, K., and M. Stiemerling, "Third-Party ALTO Server Discovery (3pdisc)", draft-kist-alto-3pdisc-05 (work in progress), January 2014. [RFC5693] Seedorf, J. and E. Burger, "Application-Layer Traffic Optimization (ALTO) Problem Statement", RFC 5693, October 2009. [RFC6708] Kiesel, S., Previdi, S., Stiemerling, M., Woundy, R., and Y. Yang, "Application-Layer Traffic Optimization (ALTO) Requirements", RFC 6708, September 2012. [RFC7216] Thomson, M. and R. Bellis, "Location Information Server (LIS) Discovery Using IP Addresses and Reverse DNS", RFC 7216, April 2014. Kiesel & Stiemerling Expires January 5, 2015 [Page 14] Internet-Draft ALTO Cross-Domain Server Discovery July 2014 [RFC7285] Alimi, R., Penno, R., and Y. Yang, "Application-Layer Traffic Optimization (ALTO) Protocol", RFC 7285, June 2014. [RFC7286] Kiesel, S., Stiemerling, M., Schwan, N., Scharf, M., and H. Song, "Application-Layer Traffic Optimization (ALTO) Server Discovery", RFC 7286, June 2014. Kiesel & Stiemerling Expires January 5, 2015 [Page 15] Internet-Draft ALTO Cross-Domain Server Discovery July 2014 Appendix A. ALTO and Tracker-based Peer-to-Peer Applications The ALTO protocol specification [RFC7285] details how an ALTO client can query an ALTO server for guiding information and receive the corresponding replies. However, in the considered scenario of a tracker-based P2P application, there are two fundamentally different possibilities where to place the ALTO client: 1. ALTO client in the resource consumer ("peer") 2. ALTO client in the resource directory ("tracker") In the following, both scenarios are compared in order to explain the need for third-party ALTO queries. In the first scenario (see Figure 2), the resource consumer queries the resource directory for the desired resource (F1). The resource directory returns a list of potential resource providers without considering ALTO (F2). It is then the duty of the resource consumer to invoke ALTO (F3/F4), in order to solicit guidance regarding this list. In the second scenario (see Figure 4), the resource directory has an embedded ALTO client, which we will refer to as 3PAC (Third-Party ALTO Client) in this document. After receiving a query for a given resource (F1) the resource directory invokes the 3PAC to evaluate all resource providers it knows (F2/F3). Then it returns a, possibly shortened, list containing the "best" resource providers to the resource consumer (F4). Kiesel & Stiemerling Expires January 5, 2015 [Page 16] Internet-Draft ALTO Cross-Domain Server Discovery July 2014 ............................. ............................. : Tracker : : Peer : : ______ : : : : +-______-+ : : k good : : | | +--------+ : P2P App. : +--------+ peers +------+ : : | N | | random | : Protocol : | ALTO- |------>| data | : : | known |====>| pre- |*************>| biased | | ex- | : : | peers, | | selec- | : transmit : | peer |------>| cha- | : : | M good | | tion | : n peer : | select | n-k | nge | : : +-______-+ +--------+ : IDs : +--------+ bad p.+------+ : :...........................: :.....^.....................: | | ALTO | client protocol __|___ +-______-+ | | | ALTO | | server | +-______-+ Figure 1: Tracker-based P2P Application with random peer preselection Peer w. ALTO cli. Tracker ALTO Server --------+-------- --------+-------- --------+-------- | F1 Tracker query | | |======================>| | | F2 Tracker reply | | |<======================| | | F3 ALTO client protocol query | |---------------------------------------------->| | F4 ALTO client protocol reply | |<----------------------------------------------| | | | ==== Application protocol (i.e., tracker-based P2P app protocol) ---- ALTO client protocol Figure 2: Basic message sequence chart for resource consumer- initiated ALTO query Kiesel & Stiemerling Expires January 5, 2015 [Page 17] Internet-Draft ALTO Cross-Domain Server Discovery July 2014 ............................. ............................. : Tracker : : Peer : : ______ : : : : +-______-+ : : : : | | +--------+ : P2P App. : k good peers & +------+ : : | N | | ALTO- | : Protocol : n-k bad peers | data | : : | known |====>| biased |******************************>| ex- | : : | peers, | | peer | : transmit : | cha- | : : | M good | | select | : n peer : | nge | : : +-______-+ +--------+ : IDs : +------+ : :.....................^.....: :...........................: | | ALTO | client protocol __|___ +-______-+ | | | ALTO | | server | +-______-+ Figure 3: Tracker-based P2P Application with ALTO client in tracker Peer Tracker w. 3PAC ALTO Server --------+-------- --------+-------- --------+-------- | F1 Tracker query | | |======================>| | | | F2 ALTO cli. p. query | | |---------------------->| | | F3 ALTO cli. p. reply | | |<----------------------| | F4 Tracker reply | | |<======================| | | | | ==== Application protocol (i.e., tracker-based P2P app protocol) ---- ALTO client protocol Figure 4: Basic message sequence chart for third-party ALTO query Note: the message sequences depicted in Figure 2 and Figure 4 may occur both in the target-aware and the target-independent query mode (c.f. [RFC6708]). In the target-independent query mode no message exchange with the ALTO server might be needed after the tracker query, because the candidate resource providers could be evaluated using a locally cached "map", which has been retrieved from the ALTO server some time ago. Kiesel & Stiemerling Expires January 5, 2015 [Page 18] Internet-Draft ALTO Cross-Domain Server Discovery July 2014 The problem with the first approach is, that while the resource directory might know thousands of peers taking part in a swarm, the list returned to the resource consumer is usually shortened for efficiency reasons. Therefore, the "best" (in the sense of ALTO) potential resource providers might not be contained in that list anymore, even before ALTO can consider them. For illustration, consider a simple model of a swarm, in which all peers fall into one of only two categories: assume that there are "good" ("good" in the sense of ALTO's better-than-random peer selection, based on an arbitrary desired rating criterion) and "bad' peers only. Having more different categories makes the maths more complex but does not change anything to the basic outcome of this analysis. Assume that the swarm has a total number of N peers, out of which are M "good" and N-M "bad" peers, which are all known to the tracker. A new peer wants to join the swarm and therefore asks the tracker for a list of peers. If, according to the first approach, the tracker randomly picks n peers from the N known peers, the result can be described with the hypergeometric distribution. The probability that the tracker reply contains exactly k "good" peers (and n-k "bad" peers) is: / m \ / N - m \ \ k / \ n - k / P(X=k) = --------------------- / N \ \ n / / n \ n! with \ k / = ----------- and n! = n * (n-1) * (n-2) * .. * 1 k! (n-k)! The probability that the reply contains at most k "good" peers is: P(X<=k)=P(X=0)+P(X=1)+..+P(X=k). For example, consider a swarm with N=10,000 peers known to the tracker, out of which M=100 are "good" peers. If the tracker randomly selects n=100 peers, the formula yields for the reply: P(X=0)=36%, P(X<=4)=99%. That is, with a probability of approx. 36% this list does not contain a single "good" peer, and with 99% probability there are only four or less of the "good" peers on the list. Processing this list with the guiding ALTO information will ensure that the few favorable peers are ranked to the top of the Kiesel & Stiemerling Expires January 5, 2015 [Page 19] Internet-Draft ALTO Cross-Domain Server Discovery July 2014 list; however, the benefit is rather limited as the number of favorable peers in the list is just too small. Much better traffic optimization could be achieved if the tracker would evaluate all known peers using ALTO, and return a list of 100 peers afterwards. This list would then include a significantly higher fraction of "good" peers. (Note, that if the tracker returned "good" peers only, there might be a risk that the swarm might disconnect and split into several disjunct partitions. However, finding the right mix of ALTO-biased and random peer selection is out of the scope of this document.) Therefore, from an overall optimization perspective, the second scenario with the ALTO client embedded in the resource directory is advantageous, because it is ensured that the addresses of the "best" resource providers are actually delivered to the resource consumer. An architectural implication of this insight is that the ALTO server discovery procedures must support third-party discovery. That is, as the tracker issues ALTO queries on behalf of the peer which contacted the tracker, the tracker must be able to discover an ALTO server that can give guidance suitable for that respective peer. Kiesel & Stiemerling Expires January 5, 2015 [Page 20] Internet-Draft ALTO Cross-Domain Server Discovery July 2014 Appendix B. Contributors List and Acknowledgments The initial version of this document was co-authored by Marco Tomsu (Alcatel-Lucent). This document borrows some text from [RFC7286], as it was historically part of that memo. Special thanks to Michael Scharf and Nico Schwan. Kiesel & Stiemerling Expires January 5, 2015 [Page 21] Internet-Draft ALTO Cross-Domain Server Discovery July 2014 Authors' Addresses Sebastian Kiesel University of Stuttgart Information Center Allmandring 30 Stuttgart 70550 Germany Email: ietf-alto@skiesel.de URI: http://www.rus.uni-stuttgart.de/nks/ Martin Stiemerling University of Applied Sciences Darmstadt, Computer Science Dept. Haardtring 100 Darmstadt 64295 Germany Phone: +49 6151 16 7938 Email: mls.ietf@gmail.com URI: http://ietf.stiemerling.org Kiesel & Stiemerling Expires January 5, 2015 [Page 22]