RFC 6972
Problem Statement and Requirements of the Peer-to-Peer Streaming Protocol (PPSP)
Problem Statement and Requirements of the Peer-to-Peer Streaming Protocol (PPSP) 🎀
Okay bestie, this one is RFC 6972 📄✨ It's marked Informational, which means it's here to explain things rather than set a standard. It was written by Y. Zhang (Coolpad) and N. Zong (Huawei Technologies) and dated July 2013. Its ISSN is 2070-1721. It doesn't obsolete or update any other RFCs, so it's a solo main character 💅
Abstract
Okay so P2P streaming is getting SO popular on the Internet, and most of the systems out there are running on proprietary protocols 👀 This doc points out the problems that come with proprietary protocols 🙄, proposes building the Peer-to-Peer Streaming Protocol (PPSP) 🌈 (which covers both the tracker and peer protocols), and talks about PPSP's scope, requirements, and use cases. It's giving "let's standardize this" energy 💖
Status of This Memo
Real talk first: this doc is not an Internet Standards Track specification. It's published just to inform people 🌸
It's a product of the Internet Engineering Task Force (IETF) and represents the consensus of the IETF community. It went through public review and got approved for publication by the Internet Engineering Steering Group (IESG). Not every IESG-approved doc is a candidate for any level of Internet Standard, though. Check out Section 2 of RFC 5741 for the details 🧐
For the current status, any errata, and how to send feedback, head to http://www.rfc-editor.org/info/rfc6972 🔗
Copyright Notice
Copyright (c) 2013 IETF Trust and the persons identified as the document authors. All rights reserved 🦋
This doc is subject to BCP 78 and the IETF Trust's Legal Provisions Relating to IETF Documents (http://trustee.ietf.org/license-info) as they were on the publication date. Please actually read these, because they lay out your rights and restrictions for this doc 📜 If you pull Code Components out of this document, they must include Simplified BSD License text as described in Section 4.e of the Trust Legal Provisions, and they come with no warranty, as described in the Simplified BSD License 🫶
Table of Contents
1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . 3
1.1. Backgrounds . . . . . . . . . . . . . . . . . . . . . . . 3
1.2. Requirements Language . . . . . . . . . . . . . . . . . . 3
2. Terminology and Concepts . . . . . . . . . . . . . . . . . . . 3
3. Problem Statement . . . . . . . . . . . . . . . . . . . . . . 5
3.1. Heterogeneous P2P Traffic and P2P Cache Deployment . . . . 5
3.2. QoS Issue and CDN Deployment . . . . . . . . . . . . . . . 5
3.3. Extended Applicability in Mobile and Wireless
Environments . . . . . . . . . . . . . . . . . . . . . . . 6
4. Tasks of PPSP: Standard Peer-to-Peer Streaming Protocols . . . 7
4.1. Tasks and Design Issues of the Tracker Protocol . . . . . 8
4.2. Tasks and Design Issues of the Peer Protocol . . . . . . . 9
5. Use Cases of PPSP . . . . . . . . . . . . . . . . . . . . . . 9
5.1. Worldwide Provision of Live/VoD Streaming . . . . . . . . 9
5.2. Enabling CDN for P2P VoD Streaming . . . . . . . . . . . . 11
5.3. Cross-Screen Streaming . . . . . . . . . . . . . . . . . . 12
5.4. Cache Service Supporting P2P Streaming . . . . . . . . . . 13
5.5. Proxy Service Supporting P2P Streaming . . . . . . . . . . 14
5.5.1. Home Networking Scenario . . . . . . . . . . . . . . . 14
5.5.2. Browser-Based HTTP Streaming . . . . . . . . . . . . . 14
6. Requirements of PPSP . . . . . . . . . . . . . . . . . . . . . 15
6.1. Basic Requirements . . . . . . . . . . . . . . . . . . . . 15
6.2. Operational and Management Requirements . . . . . . . . . 15
6.2.1. Operational Considerations . . . . . . . . . . . . . . 16
6.2.2. Management Considerations . . . . . . . . . . . . . . 17
6.3. PPSP Tracker Protocol Requirements . . . . . . . . . . . . 17
6.4. PPSP Peer Protocol Requirements . . . . . . . . . . . . . 18
7. Security Considerations . . . . . . . . . . . . . . . . . . . 19
8. Acknowledgements . . . . . . . . . . . . . . . . . . . . . . . 21
9. References . . . . . . . . . . . . . . . . . . . . . . . . . . 21
9.1. Normative References . . . . . . . . . . . . . . . . . . . 21
9.2. Informative References . . . . . . . . . . . . . . . . . . 211. Introduction ✨
1.1. Backgrounds
Streaming traffic is one of the biggest and fastest-growing kinds of traffic on the Internet [Cisco] 📈 P2P streaming is a huge part of why. It's really scalable and fault tolerant, since there isn't one single point of failure 💪 So P2P streaming apps can push large-scale live and video-on-demand (VoD) streams out to a massive audience with only a handful of servers 🍓 More and more providers are joining the P2P streaming scene too. For example, Content Distribution Networks (CDN) providers have started using P2P tech to deliver their streaming content 🌈
Since P2P streaming is getting built into the global content delivery infrastructure, the ecosystem needs an open, standard streaming signaling protocol suite 🙌 Almost every existing system uses proprietary protocols, and that's a mess. Having several proprietary protocols that do similar jobs means people keep rebuilding the same stuff for every new system 😩 And the lock-in makes it really hard to integrate with other players, like CDNs. Like, when existing caches and CDN systems try to support P2P streaming, proprietary protocols can make things way more complicated, since they have to deal with lots of different P2P streaming apps 🫠
So in this doc, we're proposing an open P2P Streaming Protocol, abbreviated as PPSP, to standardize the signaling operations in P2P streaming systems and solve those problems above 💖
1.2. Requirements Language
The key words "MUST" and "MUST NOT" in this document are to be interpreted as described in RFC 2119 [RFC2119] and indicate requirement levels for compliant implementations 📏
2. Terminology and Concepts 📖
Here's the vocab you need to know, bestie 🧚
- CHUNK: A CHUNK is a basic unit of data organized in P2P streaming for storage, scheduling, advertisement, and exchange among peers [VoD]. A CHUNK size varies from several KBs to several MBs in different systems. In the case of the MB size CHUNK scenario, a sub-CHUNK structure named piece is often defined to fit in a single transmitted packet. A streaming system may use different granularities for different usage, e.g., using CHUNKs during data exchange and using a larger unit such as a set of CHUNKs during advertisement.
- CHUNK ID: The identifier of a CHUNK in a content stream.
- CLIENT: A CLIENT refers to a participant in a P2P streaming system that only receives streaming content. In some cases, a node not having enough computing and storage capabilities will act as a CLIENT. Such a node can be viewed as a specific type of PEER.
- CONTENT DISTRIBUTION NETWORK (CDN): A CDN is a collection of nodes that are deployed, in general, at the network edge, like Points of Presence (POP) or Data Centers (DC), and store content provided by the original content servers. Typically, CDN nodes serve content to the users located nearby topologically.
- LIVE STREAMING: LIVE STREAMING refers to a scenario where all the audiences receive streaming content for the same ongoing event. It is desired that the lags between the play points of the audiences and streaming source be small.
- P2P CACHE: A P2P CACHE refers to a network entity that caches P2P traffic in the network and, either transparently or explicitly, streams content to other PEERs.
- PEER: A PEER refers to a participant in a P2P streaming system that not only receives streaming content, but also caches and streams streaming content to other participants.
- PEER LIST: A list of PEERs that are in the same SWARM maintained by the TRACKER. A PEER can fetch the PEER LIST of a SWARM from the TRACKER or from other PEERs in order to know which PEERs have the required streaming content.
- PEER ID: The identifier of a PEER such that other PEERs, or the TRACKER, can refer to the PEER by using its ID.
- PEER-TO-PEER STREAMING PROTOCOL (PPSP): PPSPs refer to the primary signaling protocols among various P2P streaming system components, including the TRACKER and the PEER.
- TRACKER: A TRACKER refers to a directory service that maintains a list of PEERs participating in a specific audio/video channel or in the distribution of a streaming file. Also, the TRACKER answers PEER LIST queries received from PEERs. The TRACKER is a logical component that can be centralized or distributed.
- VIDEO ON DEMAND (VoD): VIDEO ON DEMAND refers to a scenario in which different audiences may watch different parts of the same recorded streaming with downloaded content.
- SWARM: A SWARM refers to a group of PEERs that exchange data to distribute CHUNKs of the same content (e.g., video/audio program, digital file, etc.) at a given time.
- SWARM ID: The identifier of a SWARM containing a group of PEERs sharing a common streaming content.
- SUPER-NODE: A SUPER-NODE is a special kind of PEER deployed by ISPs. This kind of PEER is more stable with higher computing, storage, and bandwidth capabilities than normal PEERs.
3. Problem Statement 🔍
Here's where we get into the problems that proprietary protocols cause for P2P streaming apps 😤
3.1. Heterogeneous P2P Traffic and P2P Cache Deployment
ISPs are dealing with a bunch of different P2P streaming apps flooding their infrastructure 🌊 That includes their backbone and their exchange/interconnection points. So ISPs use P2P caches to store content locally, which cuts down on P2P traffic. P2P caches usually work at the chunk or file granularity 🧊
But here's the catch. Web traffic is made of HTTP requests and responses, so any caching device can serve it (as long as it supports HTTP) 🙃 P2P traffic is different, because it comes from multiple P2P apps. That means ISPs have to deploy a different type of cache for each type of P2P stream 😭
Honestly, that makes both engineering and operational costs go up dramatically 💸
3.2. QoS Issue and CDN Deployment
Compared to client/server streaming, P2P streaming gets criticized a lot for its poorer QoS performance, like longer startup delay, longer seek delay, and channel switch delay 😬 A hybrid CDN/P2P setup is a good approach to fixing this [CDN-P2P] ✨
To form that hybrid P2P+CDN architecture, the CDN has to be aware of the specific P2P streaming protocol being used in the collaboration 🧠 Just like we covered in Section 3.1, proprietary P2P protocols add complexity and make CDN deployment more expensive 🥲
3.3. Extended Applicability in Mobile and Wireless Environments
Okay so mobile and wireless networks are literally everywhere now 📱✨ and everyone is streaming nonstop 💅 It's reported that the average volume of video traffic on mobile networks had risen up to 50% in the early part of 2012 [ByteMobile]. There are also multiple prior studies exploring P2P streaming in mobile and wireless networks [Mobile-Streaming1] [Mobile-Streaming2] 🦋
Ngl, though, it's really hard to directly apply current P2P streaming protocols (even if we could reuse some of the proprietary ones) in mobile and wireless networks 😭
Here are some illustrative problems:
First, P2P streaming assumes a stable Internet connection in both downlink and uplink directions, with decent capacity and peers that can run for hours. That is not really the vibe for mobile terminals 🙈 Usually the connections are unstable and expensive in terms of energy consumption and transmission (especially in the uplink direction). So to make mobile/wireless P2P streaming feasible, trackers may need more information on peers, like packet loss rate, peer battery status, and processing capability during peer selection compared to fixed peers. Unfortunately, current protocols don't convey this kind of information 💀
Second, current practices often use a "bitmap" message to exchange chunk availability. The message size is in kilobytes and gets exchanged frequently, e.g., at an interval of several seconds or less. In a mobile environment with scarce bandwidth, the message size may need to be shortened, or it may require more efficient methods for expressing and distributing chunk-availability information, which is different from wireline P2P streaming. Main character energy for the bandwidth-starved 🌸
Third, resource-constrained peers, like mobile handsets or set-top boxes (STB), have multiple systems competing for severely limited resources when using proprietary protocols. The terminal has to install different streaming client software for different usages, e.g., some for movies and others for sports. Each of these applications competes for the same set of resources, even when one of them is running in background mode. PPSP can alleviate this problem with the basic idea that one common client software with PPSP and different scheduling plug-ins is better than different client software running at the same time, in terms of memory and disk consumption 🧚💜
4. Tasks of PPSP: Standard Peer-to-Peer Streaming Protocols ✨
PPSP aims to solve the problems mentioned above by standardizing signaling protocols that support either live or VoD streaming 🎀 It supports both centralized and distributed trackers. In distributed trackers, the tracker functionality is spread across decentralized peers. In this section, the tracker is a logical concept that can be implemented in a dedicated tracker server or in peers.
The PPSP design includes a signaling protocol between trackers and peers (the PPSP "tracker protocol") and a signaling protocol among the peers (the PPSP "peer protocol"), as shown in Figure 1. Together, the two protocols let peers receive streaming content within the time constraints 👑
+------------------------------------------------+
| |
| +--------------------------------+ |
| | Tracker | |
| +--------------------------------+ |
| | ^ ^ |
|Tracker | | Tracker |Tracker |
|Protocol| | Protocol |Protocol |
| | | | |
| V | | |
| +---------+ Peer +---------+ |
| | Peer |<----------->| Peer | |
| +---------+ Protocol +---------+ |
| | ^ |
| | |Peer |
| | |Protocol |
| V | |
| +---------------+ |
| | Peer | |
| +---------------+ |
| |
| |
+------------------------------------------------+
Figure 1: PPSP System ArchitectureThe PPSP design, in general, needs to solve the following challenges 🔥
- When joining a swarm, how does a peer know which peers it should contact for content?
- After determining a set of peers, how does a peer make contact with these peers? In which manner?
- How do we choose peers with better service capabilities, and how do we collect such information from peers?
- How do we improve the efficiency of the communication, e.g., which compact on-the-wire message format and suitable underlying transport mechanism (UDP or TCP)?
- How do we improve the robustness of the system using PPSP, e.g., when the tracker fails? How do we make the tracker protocol and the peer protocol loosely coupled?
4.1. Tasks and Design Issues of the Tracker Protocol
The tracker protocol handles the initial and periodic exchange of meta-information between trackers and peers, such as a peer list and content information 📋
So the tracker protocol is best modeled as a request/response protocol between peers and trackers, and it will carry the information needed to select peers suitable for real-time/VoD streaming 💖
Special tasks for the design of the tracker protocol are listed below. This is a high-level task list. The detailed requirements on the design of the tracker protocol are explained in Section 6.
- How should a peer be globally identified? This is related to the peer ID definition but is irrelevant to how the peer ID is generated.
- How do we identify different peers, e.g., peers with public or private IP addresses, peers behind or not behind NAT, peers with IPv4 or IPv6 addresses, or peers with different properties?
- The tracker protocol must be lightweight, since a tracker may need to serve a large number of peers. This is related to the encoding issue (e.g., binary-based or text-based) and the keep-alive mechanism 🦋
- How can the tracker report an optimized peer list to serve particular content? This is related to the status statistics, with which the tracker can be aware of the peer status and content status 🧠
The PPSP tracker protocol will consider all of these issues in its design according to the requirements from both the peer and tracker perspectives, and it will also take deployment and operation perspectives into consideration 🫶
4.2. Tasks and Design Issues of the Peer Protocol
The peer protocol controls the advertising and exchange of content between peers 🌈
So the peer protocol is modeled as a gossip-like protocol with periodic exchanges of neighbor and chunk-availability information 🍓
Special tasks for the design of the peer protocol are listed below. This is a high-level task list. The detailed requirements on the design of the peer protocol are explained in Section 6.
- How is certain content globally identified and verified? Since the content can be retrieved from everywhere, how do we make sure the content exchanged between peers is authentic?
- How do we identify the chunk availability in certain content? This is related to chunk addressing and chunk-state maintenance. Considering the large number of chunks in certain content, a light-weight expression is necessary.
- How do we ensure peer protocol efficiency? As we mentioned in Section 3, the chunk-availability information exchange is quite frequent. Balancing the size and amount of information exchanged is a big challenge 😭
The PPSP peer protocol will consider all of the above issues in its design according to the requirements from the peer perspective 💜
5. Use Cases of PPSP 🎬
This section is not a to-do list for the WG. It provides details on how PPSP could be used in practice 🌸
5.1. Worldwide Provision of Live/VoD Streaming 🌍✨
okay bestie, so the content provider can totally boost live streaming coverage by bringing PPSP in between different providers 💅 this is lowkey super similar to the CDNI situation in [RFC6707] [RFC6770] 🫶
picture this: provider A (like in China) is the only one doing live streaming in provider B's coverage (like the USA) and provider C's coverage (like Europe) 🌸 without PPSP, when a user (a Chinese American in the USA, iykyk 🙈) asks the tracker (which lives in A's coverage) for the program, the tracker will usually hand back a peer list stuffed with people in China, because most viewers are in China and there are only a few in the USA. that's a total vibe-killer for the user 😭 but if we get the PPSP tracker protocol to involve B and C in the cooperative provision (see Figure 2), then even when the stream doesn't pull many viewers in the USA and Europe, the tracker in A can optimally return a peer list that includes B's and C's Super-Nodes (SN for short) so the user gets a much better experience 🦋 and B's User2 and C's User3 can swap data (availability) with those local SNs using the peer protocol 🎀
+-------------------------------------------------------------------+
| |
| +------------------+ |
| +------------>| A's Tracker |<----------+ |
| | +------------------+ | |
| Tracker| ^ ^ | |
| Protocol| Tracker| |Tracker |Tracker |
| | Protocol| |Protocol |Protocol |
| | | | | |
| | | | | |
| v v v v |
| +------+ Peer +------+ +------+ +------+ |
| | B's |<------->| B's | | C's | | C's | |
| | SN1 |Protocol | SN2 | | SN1 | | SN2 | |
| +------+ +------+ +------+ +------+ |
| ^ ^ ^ ^ |
| | | | | |
| | | Peer Protocol Peer Protocol| | |
| Peer | +-------------+ +--------------+ |Peer |
| Protocol| | | |Protocol|
| | | | | |
| | | | | |
| | | | | |
| v v v v |
| +------+ Peer +------+ +---------+ Peer +---------+ |
| | A's |<------> | B's | |A's |<------> |C's | |
| | User1|Protocol | User2| | User1 |Protocol | User3 | |
| +------+ +------+ +---------+ +---------+ |
| |
+-------------------------------------------------------------------+Figure 2: Cooperative Vendors Interaction
5.2. Enabling CDN for P2P VoD Streaming 🎬
here's Figure 3, which shows how you can enable a CDN to support P2P VoD streaming from different content providers by bringing PPSP inside CDN overlays 💖 it's basically Figure 2 again, except the middle SNs are swapped out for 3rd party CDN surrogates 🧚 the CDN nodes talk to the different streaming systems (trackers and peers included) using the same PPSP protocols, so everyone's speaking the same language ✨
+-------------------------------------------------------------------+
| |
| +-------------+ +--------------+ |
| +----->| A's Tracker | | B's Tracker |<---+ |
| | +-------------+ +--------------+ | |
| Tracker| ^ ^ ^ ^ | |
| Protocol| Tracker| |Tracker | |Tracker |Tracker |
| | Protocol| |Protocol| |Protocol |Protocol|
| | | | | | | |
| | | | | | | |
| v v | | v v |
| +------+ Peer +------+| | +------+Internal+------+ |
| | CDN |<------>| CDN || | | CDN |<-----> | CDN | |
| | Node1|Protocol| Node2|| | | Node3|Protocol| Node4| |
| +------+ +------+| | +------+ +------+ |
| ^ ^ | | ^ ^ |
| | | | | | | |
| | | Peer Protocol | | HTTP | | |
| Peer | +-------------+ | | +------+ |Peer |
| Protocol| | | | | Protocol |Protocol|
| | | +-+ | | | |
| | | | | | | |
| | | | | | | |
| v v v v v v |
| +------+ Peer +------+ +---------+ Peer +---------+ |
| | A's |<------> | A's | |B's |<------> |B's | |
| | User1|Protocol | User2| | User3 |Protocol | User4 | |
| +------+ +------+ +---------+ +---------+ |
| |
+-------------------------------------------------------------------+Figure 3: CDN Supporting P2P Streaming
the interaction between CDN nodes can be handled by either existing (maybe proprietary) protocols or the PPSP peer protocol 🤓 the peer protocol is super handy for building new CDN systems (like an operator CDN) that can support streaming at a low cost 💰
also, for compatibility reasons, CDNs can support both HTTP and P2P streaming from the users' perspective 🌈
5.3. Cross-Screen Streaming 📺📱
this one is so good 😭 PCs, STB/TVs, and mobile terminals from both fixed and mobile/wireless networks all share the streaming content together 🫶 with PPSP, peers can figure out the types of access networks, average load, and peer abilities, and they can also find out what content other peers have, even across different networks (the content availability expression might get converted between networks) as shown in Figure 4 👑
+------------------------------------------------------------------+
| |
| Tracker Protocol +---------+ Tracker Protocol |
| +-------------> | Tracker |<------------------+ |
| | +---------+ | |
| | ^ | |
| | | | |
| | | | |
| V | V |
| +------+ | +------------+ |
| | STB | Tracker Protocol |Mobile Phone| |
| +------+ | +------------+ |
| ^ | ^ |
| | | | |
| | | | |
| | V | |
| |Peer Protocol +---------+ Peer Protocol | |
| +-------------> | PC |<------------------+ |
| +---------+ |
| |
+------------------------------------------------------------------+Figure 4: Heterogeneous P2P Streaming with PPSP
this info is a huge deal for picking good peers 🔥 for example, a PC or STB is more likely to provide stable content, while a mobile peer inside a high-load cell is unlikely to get selected, which could push more load onto the base station 🥺
5.4. Cache Service Supporting P2P Streaming 🗄️💜
check out Figure 5 🦋 when peers request P2P streaming data, the cache nodes step in and intercept those requests, then go grab the frequently visited content (or part of it) on behalf of the peers 🧚 to do that, the cache node asks the tracker for a peer list, and the tracker replies with external peers in that list 🙌 after the cache nodes exchange data with those peers, the cache can also act like a peer: it reports what it caches to the tracker and serves the requesting peers inside its own network afterward ✨ this whole thing can greatly cut down inter-network traffic in lots of situations and make the user experience way better 💅
+----------------------------------------------------------------+
| |
| Tracker Protocol +---------+ |
| +----------------> | Tracker | |
| | +---------+ |
| | ^ |
| | | |
| | | Tracker Protocol |
| | | |
| | | |
| | +---------|-------------------------------------|
| | | V |
| | | +---------+ |
| | +----------|---> | Cache |<-------------------+ |
| | | | +---------+ Tracker/Peer| |
| | | Peer | Protocol | |
| | | Protocol | | |
| | | | | |
| | | | | |
| V V | V |
| +-----------+ | ISP Domain +------------+ |
| | External | | | Inside | |
| | Peer | | | Peer | |
| +-----------+ | +------------+ |
+----------------------------------------------------------------+Figure 5: Cache Service Supporting Streaming with PPSP
last bit of this one: the cache nodes don't need to update their library when new applications that support PPSP come out, and that saves a lot of cost 🤑💖
5.5. Proxy Service Supporting P2P Streaming 🔮
5.5.1. Home Networking Scenario 🏠
Okay so, for apps where the peer isn't in the same device as the Media Player (like when the peer lives in a Home Media Gateway 🏡), we can bring in a PPSP Proxy 🧚, as shown in Figure 6.
+---------------------------------------------------------------+
| |
| Tracker Protocol +--------+ |
| +----------------> | Tracker| |
| | +--------+ |
| | ^ |
| | | |
| | | Tracker Protocol |
| | | |
| | +---------|------------------------------------|
| | | V |
| | | +--------+ |
| | +----------|---> | PPSP |<------------------+ |
| | | | | Proxy | DLNA | |
| | | Peer | +--------+ Protocol | |
| | | Protocol| | |
| | | | | |
| V V | V |
| +-----------+ | Home Domain +-----------+ |
| | External | | |DLNA Pres.| |
| | Peer | | |Devices | |
| +-----------+ | +-----------+ |
+---------------------------------------------------------------+Figure 6: Proxy Service Supporting P2P Streaming
So as you can see in Figure 6, the PPSP Proxy terminates both the tracker and peer protocol 💅, which lets the legacy presentation devices access P2P streaming content. It's giving "the middle person who makes everything work" energy ✨. In Figure 6, the Digital Living Network Alliance (DLNA) protocol [DLNA] is used to talk to the presentation devices, because it's super widely deployed 🌸. Obviously, other protocols can also be used 🫶.
5.5.2. Browser-Based HTTP Streaming 🌐
P2P Plug-ins are often used in browser-based environments to stream content 🍓. With P2P plug-ins, HTTP streaming can be turned into P2P streaming 🦋. From the browser's (and therefore the user's) point of view, it's just HTTP-based streaming, but the PPSP-capable plug-in can actually speed things up by pulling streams from multiple sources/peers [P2PYoutube] 🚀. In this case, the plug-ins behave just like the proxy 👯♀️.
6. Requirements of PPSP 📋
This section enumerates the requirements that should be considered when designing PPSP 💖.
6.1. Basic Requirements 🧱
PPSP.REQ-1: Each peer MUST have a unique ID (i.e., peer ID) 🆔.
It's a basic requirement for a peer to be uniquely identified in a P2P streaming system, so that other peers or trackers can refer to the peer by ID 👀.
Note that a peer can join multiple swarms with a unique ID, or change swarm without changing its ID 🔁.
PPSP.REQ-2: The streaming content MUST be uniquely identified by a swarm ID.
A swarm refers to a group of peers sharing the same streaming content 🎬. A swarm ID uniquely identifies a swarm. The swarm ID can be used in two cases: 1) a peer requests the tracker for the peer list indexed by a swarm ID; 2) a peer tells the tracker about the swarms it belongs to 📇.
PPSP.REQ-3: The streaming content MUST be partitioned into chunks 🧩.
PPSP.REQ-4: Each chunk MUST have a unique ID (i.e., chunk ID) in the swarm.
Each chunk must have a unique ID in the swarm, so that the peer can understand which chunks are stored in which peers and which chunks are requested by other peers 🕵️♀️✨.
6.2. Operational and Management Requirements ⚙️
This section lists some operational and management requirements based on the checklist presented in Appendix A of [RFC5706] 📝.
6.2.1. Operational Considerations 🛠️
PPSP.OAM.REQ-1: PPSP MUST be sufficiently configurable.
According to the basic requirements, when setting up PPSP, a content provider should generate chunk IDs and a swarm ID for each stream of content 🎥. An original content server and tracker are configured and set up. The content provider should then publish this information, typically by creating web links 🔗.
The configuration should allow the proxy-based and end-client scenarios 🖥️.
PPSP.OAM.REQ-2: PPSP MUST implement a set of configuration parameters with default values.
PPSP.OAM.REQ-3: PPSP MUST support diagnostic operations.
Mechanisms must be supported by PPSP to verify correct operation 🔍. The PPSP tracker should collect the status of the peers, including the peer's activity and whether it obtained chunks in time, etc. Such information can be used to monitor the streaming behavior of PPSP 📊.
PPSP.OAM.REQ-4: PPSP MUST facilitate achieving quality acceptable to the streaming application.
There are basic quality requirements for streaming systems 🎞️. The setup time to receive a new streaming channel, or to switch between channels, should be reasonably small ⏱️. End-to-end delay, which is the time between content generation (e.g., a camera) and content consumption (e.g., a monitor), becomes critical in the case of live streaming 😭. This is especially true when provisioning sporting events, where an end-to-end delay of 1 minute or more is not acceptable 🏀💀.
For instance, the tracker and peer protocol can carry quality-related parameters (e.g., video quality and delay requirements) together with the priorities of these parameters 🎚️. They can also carry the measured QoS situation (e.g., performance and available uplink bandwidth) of content-providing peers 📶.
PPSP implementations may use techniques such as scalable streaming to handle bandwidth shortages without disrupting playback 🌈.
6.2.2. Management Considerations 🧑💼
PPSP.OAM.REQ-5: When management objectives need to be supported in implementations, PPSP MUST support remote management using a standard interface, as well as a basic set of management information.
Due to large-scale peer networks, the PPSP tracker service or seeders should remotely collect information from peers and expose the information via a standard interface for management purposes 🗂️. Peer information can be collected via a PPSP tracker protocol or peer protocol.
The minimum set of management objects should include swarm information such as content characteristics and rate limits 📐; tracking information such as swarm list and log events 📜; and peer information such as peer activity, chunk statistics, and log events 🧾.
PPSP.OAM.REQ-6: PPSP MUST support fault monitoring, including peer and server health, as well as the streaming behavior of peers 🚨.
Peer and server health will at least include node activity and connectivity, especially for peers behind NAT 🧱. As mentioned in PPSP.OAM.REQ-4, the streaming behavior of the peer can be learned from chunk distribution information 📈.
PPSP.OAM.REQ-7: PPSP MUST support configuration management to define the configuration parameters.
A set of configurable parameters related to chunk generation in the PPSP setup stage can be defined by content providers via a management interface to content servers 🎛️.
PPSP.OAM.REQ-8: PPSP MUST support performance management with respect to streaming performance, based on chunk distribution statistics, network load, and tracker and peer monitoring 🏎️.
PPSP.OAM.REQ-9: PPSP MUST support security management. See Section 7 of this document 🔐.
6.3. PPSP Tracker Protocol Requirements 📡
PPSP.TP.REQ-1: The tracker protocol MUST allow the peer to solicit a peer list in a swarm, generated and possibly tailored by the tracker, in a query and response manner 🙋♀️.
The tracker request message may include the requesting peer's preference parameter (e.g., preferred number of peers in the peer list) or preferred downloading bandwidth 🎯. The tracker will then be able to select an appropriate set of peers for the requesting peer according to that preference.
The tracker may also generate the peer list with the help of traffic optimization services, e.g., Application-Layer Traffic Optimization [ALTO] 🗺️.
PPSP.TP.REQ-2: The tracker protocol MUST report the peer's activity in the swarm to the tracker 📣.
PPSP.TP.REQ-3: The tracker protocol MUST take the frequency of message exchange and efficient bandwidth use into consideration when communicating chunk availability information 🧮.
For example, the chunk availability information between peer and tracker can be presented in a compact method, e.g., to express a sequence of continuous "1" more efficiently 🗜️.
PPSP.TP.REQ-4: The tracker protocol MUST have a provision for the tracker to authenticate the peer 🪪.
This ensures that only authenticated users can access the original content in the P2P streaming system 🔒✨.
6.4. PPSP Peer Protocol Requirements 💖
okay so this is where the peers actually talk to each other 🫶 the peer protocol has to cover all of this:
PPSP.PP.REQ-1: The peer protocol MUST allow the peer to solicit the chunk information from other peers in a query and response manner.
PPSP.PP.REQ-2: The chunk information exchanged between a pair of peers MUST NOT be passed to other peers, unless the chunk information is validated (e.g., preventing hearsay and DoS attacks).
PPSP.PP.REQ-3: The peer protocol MUST allow the peer to solicit an additional list of peers to that received from the tracker.
✨ ngl this one makes sense: a peer might need more peers for certain streaming content, so it's allowed to ask the peers it already knows for an extra list of peers in the same swarm 🦋
PPSP.PP.REQ-4: When used for soliciting an additional list of peers, the peer protocol MUST contain swarm-membership information of the peers that have explicitly indicated they are part of the swarm, which is verifiable by the receiver.
PPSP.PP.REQ-5: The additional list of peers MUST only contain peers that have been checked to be valid and online recently (e.g., preventing hearsay and DoS attacks).
PPSP.PP.REQ-6: The peer protocol MUST report the peer's chunk availability update.
🌸 this one is so real. buffered content changes all the time and peers join and leave the swarm constantly, so what a peer's neighbors have (the peers it knows from the tracker or from other peers) keeps shifting and has to be updated on time. the update should happen at least on demand, like when a peer needs to find more peers with certain chunks, it sends a message to some other peers in the swarm asking for a content availability update. or each peer can advertise its availability to some other peers periodically 🍓 how far the update spreads and how often it's sent should be left to the algorithms, not the protocol.
PPSP.PP.REQ-7: The peer protocol MUST take the frequency of message exchange and efficient bandwidth use into consideration when communicating chunk information.
💅 for example, chunk availability info between peers can be presented in a compact way. bandwidth is precious, bestie
PPSP.PP.REQ-8: The peer protocol MUST exchange additional information, e.g., status about the peers.
🧚 this could be stuff like info about the access link, or whether a peer is running on battery or plugged into a power supply. with that info, a peer can pick more appropriate peers for streaming ⚡ main character energy for sure
7. Security Considerations 🔥
this doc talks about the problem and requirements around P2P streaming protocols without actually specifying the protocols 🧐 but it's super important to understand the security stuff that comes with the P2P nature of the solution. the main issue is that the service relies on untrusted entities (peers) for service provisioning. so malicious peers or trackers may:
- Originate DoS attacks to the trackers by sending a large number of requests with the tracker protocol;
- Originate fake information on behalf of other peers;
- Originate fake information about chunk availability;
- Originate fake reply messages on behalf of the tracker;
- Leak private information about other peers or trackers.
we list some important security requirements for PPSP protocols below 🫶
PPSP.SEC.REQ-1: PPSP MUST support closed swarms, where the peers are authenticated or in a private network.
🔐 this makes sure only trusted peers can access the original content in the P2P streaming system. it can be achieved with security mechanisms like peer authentication and/or key management schemes. another aspect is that confidentiality of the streaming content in PPSP needs to be supported. to get there, PPSP should provide mechanisms to encrypt the data exchange among the peers.
PPSP.SEC.REQ-2: Integrity of the streaming content in PPSP MUST be supported to provide a peer with the possibility of identifying unauthentic content (undesirable modifications by other entities rather than its genuine source).
🚨 in a P2P live streaming system, a polluter can introduce corrupted chunks. each receiver plays the polluted chunks it gets from its neighbors, and since peers forward chunks to other peers, the polluted content can potentially spread through the whole P2P streaming network 💀 the PPSP protocol specs will document the expected threats, how each protocol mitigates them, and the threats and mitigations when both protocols are combined in an application. this covers user privacy and protection of the content distribution too.
PPSP.SEC.REQ-3: The security mechanisms in PPSP, such as key management and checksum distribution, MUST scale well in P2P streaming systems.
8. Acknowledgements 🎀
thank you to J. Seng, G. Camarillo, R. Yang, C. Schmidt, R. Cruz, Y. Gu, A. Bakker, and S. Previdi for contributing to many sections of this document 🫶 and thank you to C. Williams, V. Pascual, and L. Xiao for contributing to the PPSP requirements section 🌈
we'd also like to acknowledge the following people who gave review, feedback, and suggestions to this document: M. Stiemerling, D. Bryan, E. Marocco, V. Gurbani, R. Even, H. Zhang, D. Zhang, J. Lei, H. Song, X. Jiang, J. Seedorf, D. Saumitra, A. Rahman, J. Pouwelse, W. Eddy, B. Claise, D. Harrington, J. Arkko, and all the AD reviewers 👑
9. References ✨
9.1. Normative References
- [RFC2119]: Bradner, S., "Key words for use in RFCs to Indicate Requirement Levels", BCP 14, RFC 2119, March 1997.
- [RFC5706]: Harrington, D., "Guidelines for Considering Operations and Management of New Protocols and Protocol Extensions", RFC 5706, November 2009.
- [RFC6707]: Niven-Jenkins, B., Le Faucheur, F., and N. Bitar, "Content Distribution Network Interconnection (CDNI) Problem Statement", RFC 6707, September 2012.
- [RFC6770]: Bertrand, G., Stephan, E., Burbridge, T., Eardley, P., Ma, K., and G. Watson, "Use Cases for Content Delivery Network Interconnection", RFC 6770, November 2012.
9.2. Informative References
- [ALTO]: Alimi, R., Penno, R., and Y. Yang, "ALTO Protocol", Work in Progress, December 2009.
- [ByteMobile]: Bytemobile, "Mobile Video Traffic Hits Nearly 70% on Certain Networks", February 2012,
<http://www.bytemobile.com/news-events/2012/archive_230212.html>. - [CDN-P2P]: Xu, D., Kulkarni, S., Rosenberg, C., and H-K. Chai, "Analysis of a CDN-P2P Hybrid Architecture for Cost-Effective Streaming Media Distribution", Multimedia Systems, vol. 11, no. 4, pp. 383-399, 2006.
- [Cisco]: Cisco, "Cisco Visual Networking Index: Forecast and Methodology, 2012 - 2017", Visual Networking Index (VNI),
<http://www.cisco.com/en/US/solutions/collateral/ns341/ns525/ns537/ns705/ns827/ white_paper_c11-481360_ns827_Networking_Solutions_White_Paper.html>. - [DLNA]: "DLNA",
<http://www.dlna.org>. - [Mobile-Streaming1]: Noh, J., Makar, M., and B. Girod, "Streaming To Mobile Users In A Peer-to-Peer Network", MOBIMEDIA , 2009.
- [Mobile-Streaming2]: Peltotalo, J., Harju, J., Saukkoh, M., Vaatamoinen, L., Bouazizi, I., Curcio, I., and J. van Gassel, "A Real-Time Peer-to-Peer Streaming System for Mobile Networking Environment", Proceedings of the INFOCOM and Workshop on Mobile Video Delivery (MoVID '09), 2009.
- [P2PYoutube]: "Youtube Extension-Opera Add-Ons", Opera Software,
<https://addons.opera.com/en/extensions/details/p2p-youtube/>. - [VoD]: Huang, Y., Fu, T., Chiu, D-M., Lui, J., and C. Huang, "Challenges, Design and Analysis of a Large-Scale P2P-VoD System", SIGCOMM , 2008.
Authors' Addresses 🦋
Yunfei Zhang Coolpad
EMail: hishigh@gmail.com
Ning Zong Huawei Technologies
EMail: zongning@huawei.com
📦 where this page comes from
- ietf: rfc6972 2013-07
- 💖 girlified by claude-haiku-5-5 on : a language model's rewrite of the original, under its terms