Internet DRAFT - draft-ietf-decade-integration-example
draft-ietf-decade-integration-example
DECADE N. Zong, Ed.
Internet-Draft X. Chen
Intended status: Informational Z. Huang
Expires: September 13, 2012 Huawei Technologies
L. Chen
HP Labs
H. Liu
Yale University
March 12, 2012
Integration Examples of DECADE System
draft-ietf-decade-integration-example-03
Abstract
Decoupled Application Data Enroute (DECADE) system is an in-network
storage infrastructure which is still under discussion and
standardization process in IETF DECADE WG. This document presents
two detailed examples of how to integrate such in-network storage
infrastructure into peer-to-peer (P2P) applications to achieve more
efficient content distribution, and Application Layer Traffic
Optimization (ALTO) system to build a content distribution platform
for Content Providers (CPs).
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 September 13, 2012.
Copyright Notice
Copyright (c) 2012 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
Zong, et al. Expires September 13, 2012 [Page 1]
Internet-Draft Integration Examples of DECADE System March 2012
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.
This document may contain material from IETF Documents or IETF
Contributions published or made publicly available before November
10, 2008. The person(s) controlling the copyright in some of this
material may not have granted the IETF Trust the right to allow
modifications of such material outside the IETF Standards Process.
Without obtaining an adequate license from the person(s) controlling
the copyright in such materials, this document may not be modified
outside the IETF Standards Process, and derivative works of it may
not be created outside the IETF Standards Process, except to format
it for publication as an RFC or to translate it into languages other
than English.
Zong, et al. Expires September 13, 2012 [Page 2]
Internet-Draft Integration Examples of DECADE System March 2012
Table of Contents
1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . 5
2. Terminology . . . . . . . . . . . . . . . . . . . . . . . . . 6
2.1. INS Server . . . . . . . . . . . . . . . . . . . . . . . . 6
2.2. INS Client . . . . . . . . . . . . . . . . . . . . . . . . 6
2.3. INS Operations . . . . . . . . . . . . . . . . . . . . . . 6
2.4. INS System . . . . . . . . . . . . . . . . . . . . . . . . 6
2.5. INS Client API . . . . . . . . . . . . . . . . . . . . . . 6
2.6. INS-enabled Application Client . . . . . . . . . . . . . . 6
2.7. INS Service Provider . . . . . . . . . . . . . . . . . . . 6
2.8. INS Portal . . . . . . . . . . . . . . . . . . . . . . . . 6
3. INS Client API . . . . . . . . . . . . . . . . . . . . . . . . 6
4. Integration of P2P Live Streaming and INS System . . . . . . . 7
4.1. Integration Architecture . . . . . . . . . . . . . . . . . 7
4.1.1. Data Access Messages . . . . . . . . . . . . . . . . . 8
4.1.2. Control Messages . . . . . . . . . . . . . . . . . . . 8
4.1.3. Object Naming Scheme . . . . . . . . . . . . . . . . . 9
4.2. Design Considerations . . . . . . . . . . . . . . . . . . 9
4.2.1. Improve Efficiency for Each Connection . . . . . . . . 9
4.2.2. Reduce Control Latency . . . . . . . . . . . . . . . . 9
5. Integration of P2P File Sharing and INS System . . . . . . . . 10
5.1. Integration Architecture . . . . . . . . . . . . . . . . . 10
5.2. Message Flow . . . . . . . . . . . . . . . . . . . . . . . 11
6. Integration of ALTO and INS System for File Distribution . . . 12
6.1. Architecture Design . . . . . . . . . . . . . . . . . . . 13
6.2. CP Uploading Procedure . . . . . . . . . . . . . . . . . . 14
6.3. End User Downloading Procedure . . . . . . . . . . . . . . 15
7. Test Environment and Settings . . . . . . . . . . . . . . . . 16
7.1. Test Settings . . . . . . . . . . . . . . . . . . . . . . 16
7.2. Test Environment for P2P Live Streaming Example . . . . . 17
7.2.1. INS Server . . . . . . . . . . . . . . . . . . . . . . 17
7.2.2. P2P Live Streaming Client . . . . . . . . . . . . . . 17
7.2.3. Tracker . . . . . . . . . . . . . . . . . . . . . . . 17
7.2.4. Streaming Source Server . . . . . . . . . . . . . . . 17
7.2.5. Test Controller . . . . . . . . . . . . . . . . . . . 18
7.3. Test Environment for P2P File Sharing Example . . . . . . 18
7.3.1. INS Server . . . . . . . . . . . . . . . . . . . . . . 18
7.3.2. Vuze Client . . . . . . . . . . . . . . . . . . . . . 18
7.3.3. Tracker . . . . . . . . . . . . . . . . . . . . . . . 18
7.3.4. Test Controller . . . . . . . . . . . . . . . . . . . 19
7.3.5. HTTP Server . . . . . . . . . . . . . . . . . . . . . 19
7.3.6. PlanetLab Manager . . . . . . . . . . . . . . . . . . 19
7.4. Test Environment for Combined ALTO and INS File
Distribution System . . . . . . . . . . . . . . . . . . . 19
8. Performance Analysis . . . . . . . . . . . . . . . . . . . . . 19
8.1. Performance Metrics . . . . . . . . . . . . . . . . . . . 20
8.1.1. P2P Live Streaming . . . . . . . . . . . . . . . . . . 20
Zong, et al. Expires September 13, 2012 [Page 3]
Internet-Draft Integration Examples of DECADE System March 2012
8.1.2. P2P File Sharing . . . . . . . . . . . . . . . . . . . 20
8.1.3. Integration of ALTO and INS System for File
Distribution . . . . . . . . . . . . . . . . . . . . . 20
8.2. Results and Analysis . . . . . . . . . . . . . . . . . . . 20
8.2.1. P2P Live Streaming . . . . . . . . . . . . . . . . . . 20
8.2.2. P2P File Sharing . . . . . . . . . . . . . . . . . . . 21
8.2.3. Integrated ALTO and INS System for File
Distribution . . . . . . . . . . . . . . . . . . . . . 22
9. Short Conclusion . . . . . . . . . . . . . . . . . . . . . . . 22
10. Security Considerations . . . . . . . . . . . . . . . . . . . 22
11. IANA Considerations . . . . . . . . . . . . . . . . . . . . . 22
12. References . . . . . . . . . . . . . . . . . . . . . . . . . . 22
12.1. Normative References . . . . . . . . . . . . . . . . . . . 22
12.2. Informative References . . . . . . . . . . . . . . . . . . 23
Authors' Addresses . . . . . . . . . . . . . . . . . . . . . . . . 23
Zong, et al. Expires September 13, 2012 [Page 4]
Internet-Draft Integration Examples of DECADE System March 2012
1. Introduction
Decoupled Application Data Enroute (DECADE) system is an in-network
storage infrastructure which is still under discussion and
standardization process in IETF DECADE WG. We implemented such in-
network storage infrastructure to simulate DECADE system including
DECADE servers, DECADE clients and DECADE protocols [I-D.ietf-decade-
arch]. Therefore, in the whole draft, we use the terms of in-network
storage (INS) system, INS server, INS client, INS operations, etc.
This draft introduces some examples of integrating INS system with
existing applications. In our example systems, the core components
include INS server and INS-enabled application client. An INS server
stores data inside the network, and thereafter manages both the
stored data and access to that data. An INS-enabled application
client including INS client and native application client uses a set
of Application Programming Interfaces (APIs) to enable native
application client to utilize INS operations such as data get, data
put, storage status query, etc.
This draft presents two detailed examples of how to integrate INS
system into peer-to-peer (P2P) applications, i.e. live streaming and
file sharing, as well as an example integration of Application Layer
Traffic Optimization (ALTO) [I-D.ietf-alto-protocol] and INS system
to support file distribution. We firstly show how to extend native
P2P applications by designing the INS-enabled P2P clients and
describing the corresponding flows of INS-enabled data transmission.
Then we introduce the functional architecture and working flows of
integrated ALTO and INS system for file distribution of Content
Providers (CPs). Finally we illustrate the performance gain to P2P
applications and more efficient content distribution by effectively
leveraging the INS system. We only show the feasibility of
integrated ALTO and INS system without comparing with other content
distribution systems at this time. More information would be
provided after more experiments are done in the near future.
Please note that the P2P applications mentioned in this draft only
represent some cases out of a large number of P2P applications, while
the INS system itself can support a variety of other applications.
Moreover, the set of APIs used in our integration examples is an
experimental implementation, which is not standard and still under
development. The INS system described in this draft is only a
preliminary functional set of in-network storage infrastructure for
applications. It is designed to test the pros and cons of INS system
utilized by P2P applications and verify the feasibility of utilizing
INS system to support content distribution. We hope our examples
would be useful for further standard protocol design, rather than to
present a solution for standardization purpose.
Zong, et al. Expires September 13, 2012 [Page 5]
Internet-Draft Integration Examples of DECADE System March 2012
2. Terminology
The following terms will be used in this document.
2.1. INS Server
A server to simulate DECADE server defined in [I-D.ietf-decade-arch].
2.2. INS Client
A client to simulate DECADE client defined in [I-D.ietf-decade-arch].
2.3. INS Operations
A set of communications between INS server and INS client to simulate
DECADE protocols defined in [I-D.ietf-decade-arch].
2.4. INS System
A system including INS servers, INS clients, and INS operations.
2.5. INS Client API
A set of APIs to enable native application client to utilize INS
operations.
2.6. INS-enabled Application Client
An INS-enabled application client includes INS client and native
application client communicating through INS client API.
2.7. INS Service Provider
An INS service provider deploys INS system and provides INS service
to applications/end users. It can be Internet Service Provider (ISP)
or other parties.
2.8. INS Portal
A simulated portal operated by INS service provider to offer
applications/end users a portal to access (e.g. upload, download)
files stored in INS servers.
3. INS Client API
In order to simplify the integration of INS system with P2P
applications, we provide INS client API to native P2P clients for
Zong, et al. Expires September 13, 2012 [Page 6]
Internet-Draft Integration Examples of DECADE System March 2012
accomplishing INS operations such as data get, data put, etc. On top
of the INS client API, a native P2P client can develop its own
application specific control and data distribution flows.
We currently developed the following five basic interfaces.
o Get_Object: Get a data object from an INS server with an authorized
token.
o Put_Object: Store a data object into an INS server with an
authorized token.
o Delete_Object: Delete a data object in an INS server explicitly
with an authorized token. Note that a data object can also be
deleted implicitly by setting a Time-To-Live (TTL) value.
o Status_Query: Query current status of an application itself,
including listing stored data objects, resource (e.g. storage space)
usage, etc.
o Generate_Token: Generate an authorization token. The token can be
passed from one INS client to other INS clients to authorize other
INS clients to access data objects from its INS storage. In our P2P
live streaming example, the token is generated by INS client. In the
example of combining ALTO and INS system, the token is generated by
the CP.
4. Integration of P2P Live Streaming and INS System
We integrate an INS client into a P2P live streaming client in order
that P2P live streaming application can easily leverage INS system
for data transmission.
4.1. Integration Architecture
The architecture of the integration of P2P live streaming application
and INS system is shown in Figure 1.
Zong, et al. Expires September 13, 2012 [Page 7]
Internet-Draft Integration Examples of DECADE System March 2012
+------------------+ +------------------+
| INS-enabled | | INS-enabled |
|P2P Live Streaming| |P2P Live Streaming|
| Client | | Client |
|+----------------+| +---------------+ |+----------------+|
|| INS |+---| INS Server |---+| INS ||
|| Client || +---------------+ || Client ||
|| |+-----------------------+| ||
|+------+---------|| |+------+---------+|
| API | | | API | |
|+------+---------+| |+------+---------+|
|| Native Client |+-----------------------+| Native Client ||
|+----------------+| |+----------------+|
+------------------+ +------------------+
Figure 1
An INS-enabled P2P live streaming client uses INS client to
communicate with INS server and transmit data between itself and INS
server. It is also compatible with original P2P live streaming
signaling messages such as peer discovery, data availability
announcement, etc.
4.1.1. Data Access Messages
INS client API is called whenever an INS-enabled P2P live streaming
client wants to get data objects from (or put data objects into) the
INS server. Each data object transferred between the application
client and the INS server should go through the INS client. A data
object is a data transfer unit between the INS server and the
application client, whose size can be application-customized
according to the variable requirements of performance or sensitive
factors (e.g. low latency).
4.1.2. Control Messages
The control protocols used between the native P2P live streaming
clients are modified BitTorrent-like protocols. Please refer to [BT]
for the detailed description of BitTorrent protocols. Native P2P
live streaming client uses BitTorrent-like protocols for meta-data
exchanging. On the other hand, INS-enabled P2P live streaming client
adds an additional message on top of BitTorrent-like protocols for
token distribution. By exchanging the authorization tokens, the
application clients can retrieve or store data objects into or from
the INS servers.
Zong, et al. Expires September 13, 2012 [Page 8]
Internet-Draft Integration Examples of DECADE System March 2012
4.1.3. Object Naming Scheme
We use the hash of a data object's content for the name of the data
object. The name of a data object is generated and distributed by
the source streaming server in our example. INS-enabled P2P live
streaming client uses the name of the data object as the ID to
request and retrieve data.
4.2. Design Considerations
One essential objective of the integration is to improve the
performance of P2P live streaming application. In order to achieve
such goal, we have some important design considerations that would be
helpful to the future work of protocol development.
4.2.1. Improve Efficiency for Each Connection
In a native P2P system, a peer can establish tens or hundreds of
concurrent connections with other peers. On the other hand, it may
be expensive for an INS server to maintain many connections for a
large number of INS clients. Typically, each INS server may only
allocate and maintain M connections (in our examples, M=1) with each
INS client at a time. Therefore, we have the following design
considerations to improve the efficiency for each connection between
INS server and INS client to achieve satisfying data downloading
performance.
o Batch Request: In order to fully utilize the connection bandwidth
of INS server and reduce the overhead, an application client may
combine multiple requests in a single request to INS server.
o Larger Data Object: Data object size in existing P2P live streaming
application may be small and thus incur large control overhead and
low transport utilization. A larger data object may be needed to
more efficiently utilize the data connection between INS server and
INS client.
4.2.2. Reduce Control Latency
In a native P2P system, a serving peer sends data objects to the
requesting peer directly. Nevertheless, in an INS system, the
serving client typically only replies with an authorization token to
the requesting client, and then the requesting client uses this token
to fetch the data objects from the INS server. This process
introduces an additional control latency compared with the native P2P
system. It is even more serious in latency sensitive applications
such as P2P live streaming. Therefore, we need to consider how to
reduce such control latency.
Zong, et al. Expires September 13, 2012 [Page 9]
Internet-Draft Integration Examples of DECADE System March 2012
o Range Token: One way to reduce control latency is to use range
token. An INS-enabled P2P live streaming client may piggyback a
range token when announcing data availability to its neighbor
clients, indicating that all available data objects are accessible by
this range token. Then instead of requesting some specific data
object and waiting for the response, a neighbor client can use this
range token to access all available data objects in the INS server.
5. Integration of P2P File Sharing and INS System
We integrate an INS client into Vuze - a BitTorrent based file
sharing application client, to leverage INS system for data
transmission.
5.1. Integration Architecture
The architecture of the integration of Vuze and INS system is shown
in Figure 2.
+------------------+ +------------------+
| INS-enabled | | INS-enabled |
| Vuze Client | | Vuze Client |
|+----------------+| +---------------+ |+----------------+|
|| INS |+---| INS Server |---+| INS ||
|| Client || +---------------+ || Client ||
|| |+-----------------------+| ||
|+------+---------+| |+------+---------+|
| API | | | API | |
|+------+---------+| |+------+---------+|
|| Native Client |+-----------------------+| Native Client ||
|+----------------+| |+----------------+|
+------------------+ +------------------+
Figure 2
An INS-enabled Vuze client uses INS client to communicate with INS
server and transmit data between itself and INS server. It is also
compatible with original Vuze signaling messages such as peer
discovery, data availability announcement, etc.
In our design, INS client inserts itself into the Vuze client by
intercepting certain BitTorrent messages, and adjusting their
handling to send/receive data using the INS operations instead.
In our example, the file to be shared is divided into many objects,
with each object being named as "filename_author_partn" where author
is the original author of the file or the user who uploads the file,
n is the sequence number of the object. We will also support hash-
Zong, et al. Expires September 13, 2012 [Page 10]
Internet-Draft Integration Examples of DECADE System March 2012
based naming scheme in next version of our implementation.
5.2. Message Flow
In order for a better comparison, we firstly briefly show the diagram
of the native Vuze message exchange, and then show the corresponding
diagram including the INS system.
+--------+ +--------+
| Vuze | | Vuze |
| Client1| | Client2|
+--------+ +--------+
| |
| HandShake |
|<----------------------------------->|
| Azureus HandShake |
|<----------------------------------->|
| BT_BitField |
|<----------------------------------->|
| BT_Request |
|------------------------------------>|
| BT_Piece |
|<------------------------------------|
| |
Figure 3
In the above diagram, one can see that the key messages for data
sharing in native Vuze are "BT_BitField", "BT_Request" and
"BT_Piece". Vuze client1 and client2 exchange "BT_BitField" messages
to announce the available data objects to each other. If Vuze
client1 wants to get certain data object from client2, it sends a
"BT_Request" message to client2. Vuze client2 then return the
requested data object to client1 by a "BT_Piece" message. Please
refer to [Vuze] for the detailed description of Vuze messages.
Zong, et al. Expires September 13, 2012 [Page 11]
Internet-Draft Integration Examples of DECADE System March 2012
________ __________ __________ ________ _________
| Vuze | | INS | | INS | | Vuze | | INS |
|Client1| | Client1 | | Client2 | |Client2| | Server |
|_______| |_________| |_________| |_______| |_________|
| | | | |
| | HandShake | | |
|<----------|------------|---------->| |
| Azureus HandShake | |
|<----------|------------|---------->| |
| | BT_BitField| | |
|<----------|------------|---------->| |
| | BT_Request | | |
|-----------|----------->| | |
| | | | |
| | Redirect | | |
| |<-----------| | |
| | | Get Data | |
| |----------------------------------->|
| | |Data Object| |
| |<-----------------------------------|
| | | | |
| BT_Piece | | | |
|<----------| | | |
| | | | |
Figure 4
o Vuze client1 sends a "BT_Request" message to Vuze client2 to
request a data object as usual.
o INS client2 embedded in Vuze client2 intercepts the incoming
"BT_Request" message and then replies with a "Redirect" message which
includes INS server's address and authorization token.
o INS client1 receives the "Redirect" message and then sends an INS
message "Get Data" to the INS server to request the data object.
o INS server receives the "Get Data" message and sends the requested
data object back to INS client1 after the token check.
o INS client1 encapsulates the received data object into a "BT_Piece"
message and sends to Vuze client1.
6. Integration of ALTO and INS System for File Distribution
The objective of ALTO service is to give guidance to applications/end
users about which content servers to select in order to optimize the
Zong, et al. Expires September 13, 2012 [Page 12]
Internet-Draft Integration Examples of DECADE System March 2012
content downloading performance in an ISP network-friendly way (e.g.
reducing bandwidth consumption). The core component of ALTO service
is called ALTO server which generates the guidance based on ISP
network information. The ALTO protocol conveys such guidance from
the ALTO server to the applications/end users. The detailed
description of ALTO protocol can be found in [I-D.ietf-alto-
protocol].
In this example, we integrate ALTO and INS system to build a content
distribution platform for CPs.
6.1. Architecture Design
The integrated ALTO and INS system allows CPs to upload files to INS
servers, and end users to download files from optimal INS servers
suggested by ALTO service. Specifically, three key components are
developed as follow.
o INS Servers: operated by an INS service provider to store files
from CPs.
o INS Portal: operated by an INS service provider to 1) offer CPs a
portal site to upload files; 2) provide ALTO service to direct end
users to optimal INS servers to download files.
o CP Portal: operated by a CP to publish the URLs of the uploaded
files for end user downloading.
The architecture is as follow.
Zong, et al. Expires September 13, 2012 [Page 13]
Internet-Draft Integration Examples of DECADE System March 2012
__________ __________
| End User | | End User |
|__________| |__________|
\ /
\ _____________ /
| CP Portal |
|_____________|
|
______________________|______________________
| INS ______|_____ | +--------+
| Service | INS | | | ALTO |
| Provider | Portal |----------------+---| Server |
| /|____________| \ | +--------+
| / | | \ |
| ________/ _____|__ _|______ \________ |
|| INS | | INS | | INS | | INS | |
||Server1 | |Server2 | |Server3 | |Servern | |
||________| |________| |________| |________| |
|____________________________________________|
Figure 5
6.2. CP Uploading Procedure
CP uploads the files into INS servers first, then gets the URLs of
the uploaded files and publishes the URLs on the CP portal for end
user downloading. The flow is shown below.
_________ _________ _________
| | | INS | | INS |
| CP | | Portal | | Server |
|_________| |_________| |_________|
| | |
| HTTP_POST | |
|------------------>| |
| | Put Data |
| |----------------->|
| | Response |
| |<-----------------|
| URLs | |
|<------------------| |
| | |
Figure 6
o CP uploads the file to the INS portal site via HTTP_POST message.
o INS portal distributes the file to the dedicated INS severs using
INS message "Put Data". Note that the data distribution policies
Zong, et al. Expires September 13, 2012 [Page 14]
Internet-Draft Integration Examples of DECADE System March 2012
(e.g. how many copies of the data to which INS servers) can be
specified by CP. The dedicated INS servers can be also decided by
the INS service provider based on policies or system status (e.g.
INS server load). These issues are out of the scope of this draft.
In our example, the data stored in INS server is divided into many
objects, with each object being named as "filename_CPname_partn"
where CPname is the name of the CP who uploads the file, n is the
sequence number of the object. We will also support hash-based
naming scheme in next version of our implementation.
o When the file is uploaded successfully, CP portal will list the
URLs of the file for end use downloading.
6.3. End User Downloading Procedure
End users can visit the CP portal web pages and click the URLs for
downloading the desired files. The flow is shown below.
_________ ____________ _________ _________ _________
| | | | | INS | | ALTO | | INS |
| End User| | CP Portal | | Portal | | Server | | Server |
|_________| |____________| |_________| |_________| |_________|
| | | | |
| HTTP_Get | | | |
|------------->| | | |
| Token | | | |
|<-------------| | | |
| | | | |
| HTTP_Get | | |
|------------------------------>| | |
| | | ALTO Req | |
| | |------------>| |
| | | ALTO Resp | |
| | |<------------| |
| Optimal INS Server address | | |
|<------------------------------| | |
| | | | |
| | Get Data | |
|---------------------------------------------------------->|
| | | | |
| | Data Object | |
|<----------------------------------------------------------|
| | | | |
Figure 7
o End user visits CP portal web page, and finds the URLs for the
desired file.
Zong, et al. Expires September 13, 2012 [Page 15]
Internet-Draft Integration Examples of DECADE System March 2012
o End user clicks the hyper link, CP portal returns token to the end
user and redirects the end user to INS portal via HTTP_Get message.
o INS portal communicates with ALTO server to find the optimal INS
server storing the requested file.
o INS portal returns the optimal INS server address to the end user.
o End user connects to the optimal INS server to get data via INS
message "Get Data" after the token check.
7. Test Environment and Settings
We conduct some tests to show the results of our integration
examples. For a better performance comparison, we ran experiments
(i.e. INS integrated P2P application v.s. native P2P application) in
the same environment using the same settings.
7.1. Test Settings
Our tests ran on a wide-spread area and diverse platforms, including
a famous commercial cloud platform - Amazon EC2 [EC2] and a well
known test-bed - PlanetLab [PL]. The experimental settings are as
follows.
o Amazon EC2: We setup INS servers in Amazon EC2 cloud, including
four regions around the world - US east, US west, Europe and Asia.
o PlanetLab: We ran our P2P live streaming clients and P2P file
sharing clients (both INS-enabled and native clients) on PlanetLab on
a wild-spread area.
o Flash-crowd: Flash-crowd is an important scenario in P2P live
streaming system due to the live nature, i.e. a large number of users
join the live channel during the startup period of the event.
Therefore, we conduct experiments to test the system performance for
flash-crowd in our P2P live streaming example.
o Total supply bandwidth: Total supply bandwidth is the sum of the
capacity of bandwidth used to serve the streaming/file content, from
both servers (including source servers and INS servers) and the P2P
clients. For a fair comparison, we set the total supply bandwidth to
be the same in both tests of native and INS-enabled P2P applications.
Zong, et al. Expires September 13, 2012 [Page 16]
Internet-Draft Integration Examples of DECADE System March 2012
7.2. Test Environment for P2P Live Streaming Example
In the tests, we have some functional components running in different
platforms, including INS servers, P2P live streaming clients (INS-
enabled or native), native P2P live streaming tracker, streaming
source server and test controller, as shown in Figure 8.
+------------+ +------------+
| INS |----| INS |
| Server | | Server |
+-----+------+ +------+-----+ Amazon EC2
______________________|__________________|_______
| |
+-----+------+ +------+-----+
| Streaming |----| Streaming |
| Client |\ /| Client |
+------+-----+ \/ +------+-----+ PlanetLab
_______________________|_______/\________|_______
| / \ |
+--------------+ +------+-----+ +------+-----+
| Streaming | | | | Test |
| Source Server| | Tracker | | Controller |
+--------------+ +------------+ +------------+ Yale Lab
Figure 8
7.2.1. INS Server
INS servers ran on Amazon EC2.
7.2.2. P2P Live Streaming Client
Both INS-enabled and native P2P live streaming clients ran on
PlanetLab. Each INS-enabled P2P live streaming client connects to
the closest INS server according to its geo-location distance to the
INS servers. INS-enabled P2P live streaming clients use their INS
servers to upload streaming content to neighbor clients.
7.2.3. Tracker
A native P2P live streaming tracker ran at Yale's laboratory and
served both INS-enabled and native P2P live streaming clients during
the test.
7.2.4. Streaming Source Server
A streaming source server ran at Yale's laboratory and served both
INS-enabled and native P2P live streaming clients during the test.
Zong, et al. Expires September 13, 2012 [Page 17]
Internet-Draft Integration Examples of DECADE System March 2012
7.2.5. Test Controller
Test controller is a manager to control all machines' behaviors in
both Amazon EC2 and PlanetLab during the test.
7.3. Test Environment for P2P File Sharing Example
Functional components include Vuze client (with and without INS
client), INS servers, native Vuze tracker, HTTP server, PlanetLab
manager and test controller, as shown in Figure 9.
+-------------+ +-------------+
| | | |
|INS Server | ... |INS Server |
| | | |
+-------------+ +-------------+
/ \ \
/ \ \
/ \ \
+-------------+ +-------------+ +-------------+ +-----------+
| Vuze | | Vuze | | Vuze | | |
| Client | | Client | ... | Client |--| Tracker |
+-------------+ +-------------+ +-------------+ +-----------+
\ | /
\ | /
\ | /
+-------------+ +-------------+ +-------------+
| PlanetLab | | Test | | |
| Manager | | Controller | | HTTP Server |
+-------------+ +-------------+ +-------------+
Figure 9
7.3.1. INS Server
INS servers ran on Amazon EC2.
7.3.2. Vuze Client
Both INS-enabled and native Vuze clients ran on PlanetLab. INS
client embedded in Vuze client was automatically loaded and ran after
Vuze client start up. Vuze clients were divided into one seeding
client and multiple leeches. The seeding client ran at a Window 2003
server.
7.3.3. Tracker
Vuze client provides tracker capability, so we did not deploy our own
tracker. Tracker was enabled when making a BitTorrent file. The
Zong, et al. Expires September 13, 2012 [Page 18]
Internet-Draft Integration Examples of DECADE System March 2012
seeding client was also a tracker in our test.
7.3.4. Test Controller
Similar to the test controller in P2P live streaming case, the test
controller in Vuze example can also control all machines' behaviors
in Amazon EC2 and PlanetLab. For example, it lists all the Vuze
clients via GUI and controls them to download a specific BitTorrent
file. It ran at the same Window 2003 server with the seeding client.
7.3.5. HTTP Server
BitTorrent file was put in the HTTP server and the leeches retrieved
the BitTorrent file from the HTTP server after receiving the
downloading command from the test controller. We used Apache Tomcat
for HTTP server.
7.3.6. PlanetLab Manager
PlanetLab manager is a tool developed by University of Washington.
It presents a simple GUI to control PlanetLab nodes and perform
common tasks such as: 1) selecting nodes for your slice; 2) choosing
nodes for your experiment based on the information about the nodes;
3) reliably deploying you experiment files; 4) executing commands on
every node in parallel; 5) monitoring the progress of the experiment
as a whole, as well as viewing console output from the nodes.
7.4. Test Environment for Combined ALTO and INS File Distribution
System
For the integration of ALTO and INS systems for supporting file
distribution of CPs, we built 6 Linux virtual machines (VMs) with
Fedora13 operating system. ALTO server, INS portal, CP portal and
two INS servers ran on these VMs. Each VM has 4G CPU, 2G Memory and
10G Disk. CP uploaded files to the INS server via INS portal. End
user can choose desired file through the CP portal, and download it
from the optimal INS server chosen by the INS portal using ALTO
service.
8. Performance Analysis
We illustrate the performance gain to P2P applications and more
efficient content distribution by effectively leveraging the INS
system. Note that for the example of integrating ALTO and INS
systems to support file distribution of CPs, we only show the
feasibility of such integration without comparing the performance of
our implementation with other content distribution systems.
Zong, et al. Expires September 13, 2012 [Page 19]
Internet-Draft Integration Examples of DECADE System March 2012
8.1. Performance Metrics
8.1.1. P2P Live Streaming
To measure the performance of a P2P live streaming application, we
mainly employed the following four metrics.
o Startup delay: The duration from a peer joins the streaming channel
to the moment it starts to play.
o Piece missed rate: The number of pieces a peer loses when playing
over the total number of pieces.
o Freeze times: The number of times a peer re-buffers during playing.
o Average peer uploading rate: Average uploading bandwidth of a peer.
8.1.2. P2P File Sharing
To measure the performance of a P2P file sharing application, we
mainly employed the following three metrics.
o Download traffic: The total amount of traffic (MByte) representing
the network downlink resource usage.
o Upload traffic: The total amount of traffic (MByte) representing
the network uplink resource usage.
o Network resource efficiency: The ratio of P2P system download rate
to the total network (downlink) bandwidth.
8.1.3. Integration of ALTO and INS System for File Distribution
We only consider some common capacity metrics for content
distribution system, i.e. the bandwidth usage of each INS server, and
the total online users supported by each INS server at this time.
More comprehensive metrics would be provided after more experiments
are done in the near future.
8.2. Results and Analysis
8.2.1. P2P Live Streaming
o Startup delay: In the test, INS-enabled P2P live streaming clients
startup around 35~40 seconds and some of them startup around 10
seconds. Native P2P live streaming clients startup around 110~120
seconds and less than 20% of them startup within 100 seconds.
Zong, et al. Expires September 13, 2012 [Page 20]
Internet-Draft Integration Examples of DECADE System March 2012
o Piece missed rate: In the test, both INS-enabled P2P live streaming
clients and native P2P live streaming clients achieved a good
performance in piece missed rate. Only about 0.02% of total pieces
missed in both cases.
o Freeze times: In the test, native P2P live streaming clients
suffered from more freezing times than INS-enabled P2P live streaming
clients by 40%.
o Average peer uploading rate: In the test, according to our
settings, INS-enabled P2P live streaming clients had no data upload
in their "last mile" access network, while in the native P2P live
streaming system, most peers uploaded streaming data for serving
other peers. In another word, INS system can shift uploading traffic
from clients' "last mile" to in-network devices, which saves a lot of
expensive bandwidth on access links.
8.2.2. P2P File Sharing
The test result is illustrated in Figure 10. We can see that there
is very few upload traffic from the INS-enabled Vuze clients, while
in the native Vuze case, the upload traffic from Vuze clients is the
same as the download traffic. Network resource usage is thus reduced
in the "last mile" in the INS-enabled Vuze case. This result also
verifies that the INS system can shift uploading traffic from
clients' "last mile" to in-network devices.
+--------------------+--------------------+--------------------+
| | | |
| | Download Traffic | Upload Traffic |
| | | |
+--------------------+--------------------+--------------------+
| | | |
| INS-Enabled Vuze | 480MB | 12MB |
| | | |
+--------------------+--------------------+--------------------+
| | | |
| Native Vuze | 430MB | 430MB |
| | | |
+--------------------+--------------------+--------------------+
Figure 10
We also found higher network resource efficiency in the INS-enabled
Vuze case where the network resource efficiency is defined as the
ratio of P2P system download rate to the total network (downlink)
bandwidth. The test result is that the network resource efficiency
of native Vuze is 65% while that of INS-enabled Vuze is 88%.
Zong, et al. Expires September 13, 2012 [Page 21]
Internet-Draft Integration Examples of DECADE System March 2012
8.2.3. Integrated ALTO and INS System for File Distribution
Each INS server can supply the bandwidth usage of at most 94% of
network interface card (e.g. 1000M interface card server can supply
bandwidth of 940Mbps at most). Each INS server can support about 400
online users for file downloading simultaneously.
9. Short Conclusion
This document presents two examples of integrating INS system into
P2P applications (i.e. P2P live streaming and Vuze) by developing
INS client API for native P2P clients. To better adopt INS system,
we found some important design considerations including efficiency
for INS connection, control latency caused by INS operations, and
developed some mechanisms to address them. We ran some tests to show
the results of our integration examples on Amazon EC2 and PlanetLab
for deploying INS servers and clients, respectively. It can be
observed from our test results that integrating INS system into
native P2P applications could achieve performance gain to P2P
applications and more network efficient content distribution.
Note that for the example of integrating ALTO and INS system to
support file distribution of CPs, we only show the feasibility of
such integration without comparing with other content distribution
systems at this time. More information would be provided after more
experiments are done in the near future.
10. Security Considerations
The token can be passed from one INS client to other INS clients to
authorize other INS clients to access data objects from its INS
storage. Detailed mechanisms of token based authentication and
authorization can be found in [I-D.ietf-decade-arch].
11. IANA Considerations
This document does not have any IANA considerations.
12. References
12.1. Normative References
[I-D.ietf-decade-arch] Alimi, R., Yang, Y., Rahman, A., Kutscher, D.,
and H. Liu, "DECADE Architecture", draft-ietf-decade-arch-04 (work in
Zong, et al. Expires September 13, 2012 [Page 22]
Internet-Draft Integration Examples of DECADE System March 2012
progress), October 2011.
[I-D.ietf-alto-protocol] Alimi, R., Penno, R., and Y. Yang, "ALTO
Protocol", draft-ietf-alto-protocol-10 (work in progress), October
2011.
12.2. Informative References
[BT] "http://www.bittorrent.org"
[Vuze] "http://www.vuze.com"
[EC2] "http://aws.amazon.com/ec2/"
[PL] "http://www.planet-lab.org/"
Authors' Addresses
Ning Zong (editor)
Huawei Technologies
Email: zongning@huawei.com
Xiaohui Chen
Huawei Technologies
Email: risker.chen@huawei.com
Zhigang Huang
Huawei Technologies
Email: andy.huangzhigang@huawei.com
Lijiang Chen
HP Labs
Email: lijiang.chen@hp.com
Hongqiang Liu
Yale University
Email: hongqiang.liu@yale.edu
Zong, et al. Expires September 13, 2012 [Page 23]
