Docstoc

A Hierarchical Overlay Design for Peer to Peer and SIP Integration

Document Sample
A Hierarchical Overlay Design for Peer to Peer and SIP Integration Powered By Docstoc
					                                                             (IJCSIS) International Journal of Computer Science and Information Security,
                                                             Vol. 9, No. 6, June 2011




A Hierarchical Overlay Design for Peer to Peer and
                 SIP Integration
              Md. Safiqul Islam #1 , Syed Ashiqur Rahman #2 , Rezwan Ahmed ∗3 , Mahmudul Hasan #4
                      #
                          Computer Science and Engineering Department, Daffodil International University
                                                     Dhaka, Bangladesh
                                                1
                                                     safiqul@daffodilvarsity.edu.bd
                                                2
                                                    ashiq797@daffodilvarsity.edu.bd
                                                    4
                                                      mhraju@daffodilvarsity.edu.bd
                                          ∗
                                              American International University - Bangladesh
                                                           Dhaka, Bangldesh
                                                         3
                                                             a.rezwan@aiub.edu


   Abstract—Peer-to-Peer Session Initiation Protocol (P2PSIP) is             P2P can be integrated with a traditional SIP system. We also
the upcoming migration from the traditional client-server based              examine some existing approaches to identify in achieving
SIP system. Traditional centralized server based SIP system                  P2PSIP. The remainder of the paper is organized as follows.
is vulnerable to several problems like performance bottleneck,
single point of failure. So, integration of Peer-to-Peer system              Section II gives an overview of traditional SIP systems and
(P2P) with Session Initiation Protocol (SIP) will improve the per-           terminologies involved. P2P technology is demonstrated in
formance of a conventional SIP system because a P2P system is                section III. P2PSIP is introduced in section IV and section V.
highly scalable, robust, and fault tolerant due to its decentralized         In Section VI, details out the current approaches for P2PSIP.
manner and self-organization of the network. However, P2PSIP                 In Section VII, a proposal for the integration of P2PSIP is
architecture faces several challenges including trustworthiness
of peers, resource lookup delay, Network Address Translation                 introduced based on the existing approaches. Finally, Section
(NAT) traversal, etc. This paper focuses on understanding the                VIII takes an account of the conclusion.
needs of integration of P2P and SIP. It also reviews the existing
approaches to identify their advantages and shortcomings. Based                        II. S ESSION I NITIATION P ROTCOL (SIP)
on the existing approaches, it proposes a layered architecture to               SIP is an application-layer control protocol for initiating,
address the major challenges introduced by P2PSIP.
                                                                             terminating and modifying multimedia sessions (for example
                                                                             video, voice, instant messaging, online games and multimedia
                       I. I NTRODUCTION
                                                                             conferences). To establish a session, a traditional PSTN
   The session initiation protocol (SIP) is a signaling protocol             requires SS7 [3] signaling. In IP based telephony, the
[1] standardized by IETF. It is also the default standard                    signaling protocol is SIP. However, Session Description
protocol for VoIP. The majority of the VoIP development                      Protocol (SDP) [4] and Real-Time Transport Protocol (RTP)
is currently based on SIP. SIP is used to establish, modify,                 [5] should be used together with SIP to provide complete
or tear down a multimedia session. Most VoIP systems                         IP telephony system. Traditional SIP architecture uses
rely on fixed set of SIP servers for which they suffer from                   client-server architecture. SIP servers are classified into
performance bottlenecks, single point of failure, and Denial                 proxy, registrar, and redirect servers [1]. A proxy server is an
of Service (DoS) attacks. On the other hand, a Peer-to-Peer                  intermediate entity that can act as a server to accept a SIP
(P2P) system [2] is a popular technology which does not rely                 request or act as a client to forward a SIP request. Redirect
on central control and is very popular for resource sharing.                 servers performs redirection of SIP request. A registrar server
As in a P2P system there is no centralized server, such a                    accepts Register request from clients and maintains location
system has greater robustness, scalability and fault tolerance.              information in order to support mobility. SIP user agents are
If SIP can be made to work over P2P systems, it will improve                 classified into user agent client (UAC) and user agent server
the performance of traditional SIP systems and eliminate                     (UAS). A user agent client [1] is a SIP entity that creates
the problems of using centralized SIP servers. Integration                   a new request. It uses client state machinery to send that
of P2P technologies and SIP introduces several challenges                    request. A user agent server is also a SIP entity that receives
such as resource lookup delays, node heterogeneity, NAT                      SIP requests on behalf of users and responds to these requests.
traversal, peers trustworthiness which has to be addressed
before enjoying its advantages.                                                Each user agent is identified by SIP uniform resource
                                                                             identifier (URI) for instance sip:username@somedomain.com.
   This paper describes the need to integrate P2P and SIP                    In order to initiate a session with another user, the caller first
technologies and examines cost benefit of P2PSIP over the                     needs to know the SIP URI of that user. A caller can either
traditional fixed set of SIP servers. Further it discusses how                send an INVITE request to a locally configured SIP server




                                                                        94                              http://sites.google.com/site/ijcsis/
                                                                                                        ISSN 1947-5500
                                                     (IJCSIS) International Journal of Computer Science and Information Security,
                                                     Vol. 9, No. 6, June 2011


or directly send an INVITE to the IP address and port of                the prefix based identifier where each node shares a prefix
the user’s address. A user agent registers its location with the        with the data key .There is also a unique identifier for each
registrar server before initiating a session.                           key. The algorithm choose the node with the closest numeric
                                                                        value of a key to route messages. During routing, the sender
                 III. P EER - TO -P EER (P2P)
                                                                        node sends the message to the node whose identifier’s prefix
   Peer-to-peer(P2P) technology, by definition, is a mesh net-           is minimum one digit larger than the sender node. If there
work as opposed to a star network in a client/server model. In          is no existence of such a node it forwards the messages
a peer-to-peer network, all nodes act simultaneously as client          to the another node with the same prefix whose identifier
and server. Some of the advantages of P2P systems are:                  is numerically closer to the data key.The expected number
   • Scalability                                                        of routing step is O(logN). The Pastry method provides
   • Robustness                                                         more flexible mechanism than the chord method because the
   • No single point of failure                                         successor with the pastry identifiers is not so strictly defined
   There are two types of P2P systems [6], structured and               and its adjusting nodes in its routing table.
unstructured. The structured networks impose techniques to
tightly control the data placement and topology within the
                                                                           The Bamboo DHT algorithm conceptualize the namespace
network, and currently only support search by identifier.
                                                                        as a circle like Chord DHT which means the peer always
On the other hand unstructured networks rely on flooding
                                                                        located next to the peer with the largest possible Peer-ID
techniques. In many scenarios, the increased search efficiency
                                                                        [9]. Unlike Chord but like Pastry, bamboo uses prefix routing
makes structured networks preferable to the widely deployed
                                                                        to accumulate on the peer responsible for the search key. It
unstructured networks. Most widely used structured P2P
                                                                        uses the Pastry geometry where the term geometry is used to
system is Distributed Hash Table (DHT). There are several
                                                                        refer the neighbor management algorithm or independent of
flavors of DHT each with some advantages over another. It
                                                                        the routing algorithms used as well as patterns of neighbor
is very important to choose a proper DHT algorithm to have
                                                                        links in DHT. The Bamboo algorithms are more incremental
good performance of SIP running on P2P.
                                                                        than Pastry. In bandwidth-limited environments, the bamboo
                                                                        algorithm allows to continuous churn in membership as well as
   DHT is a decentralized and distributed system where all
                                                                        it allows to better acceptance of changes in large membership
the peer nodes and resources are identified by unique keys.
                                                                        in DHT.
A DHT table is introduced to provide efficient location and
retrieving operations. The most popular DHT algorithms are
Chord [7], Bamboo [8], Pastry [10], and Tapestry [11].
Most of the P2PSIP architectures use the Chord algorithm
to maintain the P2P overlay. The logical structure of                                                         N4     Predecessor
Chord is a ring shaped where each node is identified by a
numeric identity. For each peer, the peer with the nearest
lower identifier is called the predecessor and nearest higher                                                         N23
identifier is known as successor. Figure 1 illustrates a Chord
ring where node 23 is responsible for objects O6 to O22.
Each node maintains a finger table that contains information
                                                                                                             N35     Successor
of half of the nodes clockwise from that node. Each node is
responsible for objects whose associated key is in between
the node’s own id and predecessor’s ID.                                                       Fig. 1.   Chord Ring


   Like Chord, Pastry is another 2nd generation large P2P
routing network [10]. Pastry forms a self-organized, robust
                                                                           A node will search for a target node by searching the
and overlay network in the internet. The major challenge is
                                                                        node in the finger table which is nearest to the target
again to form an efficient algorithm for routing. In Pastry
                                                                        node. Since each node knows about nearby nodes, so after
network, each peer or node has a unique numeric 128-bit
                                                                        a repeated number of searches, the target node will be
identifier which is assigned randomly when a peer joins in
                                                                        discovered. Choosing the proper DHT algorithm will improve
the network . Each peer formed the overlay network on the
                                                                        the performance of P2PSIP. Table 1 shows a comparison
top of the hash table and the peer contains the table of list
                                                                        of different DHT algorithms [12]; from this list a suitable
of leaf nodes, a routing table and a neighborhood list. Leaf
                                                                        algorithm for a P2PSIP implementation could be chosen.
nodes, a routing and a neighborhood list tables organized
based on the existing nodes of the network and here we can
see the self-organization is very similar to Chord algorithm               Below is a table comparing the various features of different
except that Pastry also update its routing table. Leaf nodes            flavors of DHT algorithms: Comparison among different DHT
set contains L/2 closest nodes where as routing table contains          algorithms, adapted from [12] is shown in Table I.




                                                                   95                            http://sites.google.com/site/ijcsis/
                                                                                                 ISSN 1947-5500
                                                      (IJCSIS) International Journal of Computer Science and Information Security,
                                                      Vol. 9, No. 6, June 2011


                                                             TABLE I
                                   C OMPARISON OF DIFFERENT DHT ALGORITHM , ADAPTED        FROM   [12]

                               Chord         CAN             Pastry          Bamboo         Tapestry        Kademlia
                 Lookup        Recursive     Recursive      Recursive        Recursive      Recursive       Iterative
                 Methods       Semi-         Semi-          Semi-            Semi-          Semi-
                               Recursive     Recursive      Recursive        Recursive      Recursive
                               Iterative     Iterative      Iterative        Iterative      Iterative
                 Parallel      Not           No             Not              Yes (on It-    No              Yes
                 Lookups       Suitable                     suitable         erative)
                 Complexity    Simple        Simple         Quite com-       Quite com-     Not             Simple
                                                            plex             plex           complex
                 Bandwidth     Moderate      Moderate       High             Moderate       Quite high      Moderate
                 Consump-
                 tion
                 Node          Quite sim-    Very           Complex          Quite sim-     Complex         Simple
                 Join and      ple           simple         join             ple            join
                 Departure


   IV. P EER - TO -P EER S ESSION I NITIATION P ROTOCOL                A. Resource Lookup Delay
                          (P2PSIP)                                       Locating a peer or resources in P2PSIP networks takes much
                                                                       more time than the traditional SIP based network. Trying to
   P2PSIP is the combination of a P2P network and SIP, where           reduce this delay is significant challenge for P2PSIP.
traditional fixed set of servers are replaced by a distributed
mechanism. DHT can be used which is one of the possible                B. Network Address Table (NAT) Traversal
distributed mechanisms available. All the address of record              Most of the P2P nodes may be behind a NAT or Firewall.
to contact URI mappings is distributed among the peers                 There must be some relay in between them with a public
in the P2P overlay. Currently, the P2PSIP working group                IP address in order to establish communication with other
of the Internet Engineering Task Force (IETF) has defined               peer. This is one of the most important challenges for P2PSIP
the terminologies, concepts in [13] and use cases in [14].             network.
Moreover, this working group is trying to standardize the
P2PSIP peer protocol. P2PSIP can be implemented into two               C. Node Heterogeneity
ways: one is SIP on top of P2P and the other implements                   In order to maintain the scalability and service availability of
P2P over SIP. SIP on top of P2P uses P2P protocol to                   P2PSIP, node heterogeneity should be handled appropriately.
implement SIP location service; while the other approach               Node heterogeneity can be difference in bandwidth, CPU,
uses SIP messages to transport P2P traffic. Traditional a P2P           storage, and uptime of the peer. Security Issues and Trustwor-
node searching mechanism uses flooding mechanism to locate              thiness of peers Security of a distributed P2P communication
the node. However, to find a target node using P2PSIP for               system is another of the major challenges. Security issues
multimedia session, flooding mechanism should be avoided.               concern user identification, authentication and trustworthiness.

                                                                       D. Security Issues and Trustworthiness of Peers
        V. P INPOINTING C HALLENGES FOR P2PSIP
                                                                         Security of a distributed P2P communication system is
   In the following section, we will discuss some common               another of the major challenges. Security issues concern user
requirements to implement P2PSIP. In an IETF draft, Bryan,             identification, authentication and trustworthiness.
et al. defines a set of requirements for P2PSIP [15]. We have
taken some important point from this paper.                                             VI. E XISTING A PPROACHES
First, P2PSIP peers should be capable of performing opera-                Integration of P2P and SIP will improve the performance
tions such as joining, leaving, storing information on behalf          of traditional SIP system as P2P has several advantages.
of the overlay, or transporting messages. Secondly, the peers          There are several requirements that should be met to integrate
must provide the functions offered by traditional SIP network.         P2P and SIP. Those requirements have already been described
For example, P2PSIP should support the modification, estab-             in Section V.
lishment, and termination of multimedia sessions. Thirdly, the
implementation should not prevent the use of existing proto-             Singh and Schulzrinne propose a hybrid architecture for
cols like SSL or TLS as used in the P2P or SIP network. NAT            the integration of P2P and SIP that introduces two additional
and firewall traversal should also be supported for P2PSIP.             advantages: interoperability with existing SIP servers and
Finally, the functionality of the fixed set of centralized SIP          no maintenance cost besides P2P scalability and reliability
servers should be distributed over the peers. Some of the other        [16]. Chord was used as an underlying DHT algorithm.
challenges are described in the following sections.                    Their architecture is based on the concept of Super nodes




                                                                  96                               http://sites.google.com/site/ijcsis/
                                                                                                   ISSN 1947-5500
                                                     (IJCSIS) International Journal of Computer Science and Information Security,
                                                     Vol. 9, No. 6, June 2011


and ordinary node where super nodes are powerful nodes                  briefly.
having high bandwidth, lots of CPU power, and plenty of
memory, long uptimes and public IP address compared to                     Dhara, et al. [22] propose a layered architecture for P2PSIP
the ordinary nodes. These authors have also implemented                 that separates the P2P related issues from the underlying
a P2PSIP adaptor [17] which allows existing or new user                 voice or transport layer. The advantage of this system is
agents to connect to the P2PSIP network. NAT/Firewall                   that it allows dynamic changes of overlays based on the
detection is based on the ICE [18] algorithm, then it uses              requirements and properties of users and devices. Their paper
a super node as a relay to help ordinary nodes to establish             focuses on a layered architecture that allows the choice of
calls and participate in calls. Node heterogeneity is identified         P2P overlay based on specific parameters. The architecture is
by the super nodes and ordinary nodes. Offline messaging                 a Public Key Infrastructure PKI [23] based trust management
services are provided by combining the storage of sender and            system where SIP is the transport protocol and chord is DHT
intermediate DHT nodes [19].                                            algorithm for a P2P structure. They have mentioned that
                                                                        the device overlay will focus on NAT and firewall traversal,
  Their architecture fails to reduce call setup delay; which is         but no such implementation or methods were introduced.
higher than the traditional SIP networks.Peer trustworthiness           This paper failed to address DHT maintenance cost on
and security issues are not described in their report. The              an ordinary node and bootstrap node selection- Besides
paper does not propose any super node selection mechanism.              the above approach, Scalable Application-Layer Mobility
Node heterogeneity is introduced, but they do not describe              Protocol (SAMP) [24] deals with the session setup latency
what will happen if a more powerful node than the current               by introducing two optimization techniques: Hierarchical
super node joins the existing network. Ordinary nodes always            Registration (HR) and Two-Tier Caching (TTC). In HR, when
have to pay a high maintenance cost to maintain their Chord             the mobile node (MN) is in a foreign domain it will register
finger table and periodically send refresh messages to update            its Care of Address with an anchor SIP server instead of
their predecessors and successors.                                      home SIP server, this is to reduce session setup delay. On the
                                                                        other hand, TTC introduces two phase cache lookup where
   SOSIMPLE [20] is a P2PSIP architecture where nodes are               the first cache lookup is based on the MN’s cache and the
organized using the Chord DHT algorithm. SIP messages                   second phase lookup is based on anchor SIP server’s cache.
with a newly defined header are used to maintain the DHT,                If the target is not found in either cache then a traditional
register users, locate resources and establish sessions. Based          P2P lookup occurs. Hierarchical P2PSIP [25] is introduced
on the registration process, the SOSIMPLE architecture has              to address the connectivity problem of heterogeneous P2P
two levels of REGISTER operations. One is user registration             overlays and the overhead problem of extra SIP messages
which is the traditional use of registration and another is node        overhead when SIP is used to maintain the overlays.
registration which is for DHT operation. The SOSIMPLE
paper mentions several security and user authentication                    Another approach to P2PSIP is described in [26]. This
mechanisms such as user certificate, email verification, etc.             approach deals with the manageability of a P2P system. In
that can be implemented on their architecture. However, this            this approach, the system architecture is divided into three
paper fails to describe Node heterogeneity, Bootstrap node              layers: the signal control layer, the management layer, and
selection, Node maintenance cost for DHT operation, and                 the media transportation layer. In this architecture signaling
Resources look up delay.                                                is used to initiate the system, maintain the system topology,
                                                                        and for resource searching. The management layer deals with
   In order to alleviate the problems in [16] and [20], such as         grouping, playing, and media uploading management. Inter-
call setup delay, node heterogeneity, and high maintenance              domain registers, resource locating and DHT creation and
cost paid by the ordinary nodes, Le and Kuo [21] propose                maintenance are the functions of the SIP signal control layer.
a hierarchical and breathing overlay based network. In a                Media transport layer deals with the storage management and
hierarchical overlay nodes forms different sub-overlays based           media uploading. In Cooperative SIP (CoSIP) [27], both the
on node heterogeneity. Session setup delay is reduced by                server based and P2PSIP networking work together.
introducing two types of lookups based on knowledge of
the destination sub overlay: oriented lookup and un-oriented                      VII. L AYERED A RCHITECTURE FOR P2PSIP
lookup. In oriented lookup, a node relays its request to its
father node to establish the session. In the case of un-oriented           In this section, we propose a new P2PSIP architecture in
lookup, a father node will receive the request from a son               order to address all the challenges. Our P2PSIP architecture
node and this (and if necessary other) upper level father               is a two layered architecture based on [25]. Where the top
node will do a DHT lookup in their sub over-lays. Node                  layer consists of powerful super nodes and the bottom layers
heterogeneity is introduced due to forming a hierarchical               consist of the ordinary nodes similar to HiLO Peer and LoLO
overlay. Lower DHT maintenance cost. The paper failed to                Peer [25]. Figure 2 shows our proposed architecure for P2PSIP.
address peer trustworthiness and other security issues. The             The top overlay uses Bamboo as an underlying algorithm. The
following subsections will describe their three techniques              reason for choosing Bamboo is its parallel lookup capability.




                                                                   97                            http://sites.google.com/site/ijcsis/
                                                                                                 ISSN 1947-5500
                                                                      (IJCSIS) International Journal of Computer Science and Information Security,
                                                                      Vol. 9, No. 6, June 2011


Several challenges remain to our proposed approach. However,                          F. Peer’s Trustworthiness and Security issues
our proposed architecture has the following properties:
                                                                                        We can implement a login server module to the super node.
                                                                                      When a peer wants to join a P2P network, it will send a SIP
                                  Intra Domain                                        REGISTER message along with its public key, and then the
                                     Overlay
                                     (Chord)                                          login server will authenticate the node and provide it with
                                                                                      a signed certificate. It will help the node to provide this
     Intra Domain




                                                                                      certificate and build trust.




                                                                 Intra Domain
        (Pastry)
        Overlay




                     Inter Domain Overlay (Using Bamboo as an




                                                                    Overlay
                                                                     (CAN)
                               underlying algorithm)
                                                                                                                VIII. C ONCLUSION

                                   Intra Domain
                                                                                         This paper gave an overview of SIP and P2P and exam-
                                      Overlay                                         ines the integration of a P2P system with SIP. Moreover,
                                     (Tapestry)
                                                         Super Node
                                                                                      it describes some advantages of their integration as well
                                                                                      as challenges and implications were also discussed. Several
                                                         Ordinary Node
                                                                                      current approaches to P2PSIP were discussed along with their
                    Fig. 2.   Layered Architecture of P2PSIP                          approaches to mitigating the challenges of P2PSIP. Finally, we
                                                                                      proposed a layered architecture to address the major challenges
                                                                                      of P2PSIP. Future work will focus on implementing our
A. Node Heterogeneity                                                                 proposed architecture on a simulator.
  A joining node will contract the regional super node and
share his capabilities with the super node. If it has higher                                                        R EFERENCES
capabilities it will become the super node. So, here based on                         [1] J. Rosenberg, H. Schulzrinne, G. Camarillo, A. Johnston, J. Peterson, R.
node heterogeneity we select the super node.                                              Sparks, M. Handley, and E. Schooler. SIP: Session Initiation Protocol,
                                                                                          RFC 3261, Internet Engineering Task Force, 2002.
                                                                                      [2] D.S. Milojicic, V. Kalogeraki, R. Lukose, K. Nagaraja, J. Pruyne, B.
B. High Maintenance Cost Problem                                                          Richard, S. Rollins, and Z. Xu. Peer-to-Peer Computing, HP Laboratories
   As the ordinary node has to pay a high maintenance cost                                Palo Alto, 2002.
                                                                                      [3] Performance Technologies, Inc. SS7 Tutorial, 2006. Available at:
for the DHT, we use the concept of a breathing layer with a                               http://www.pt.com/tutorials/ss7/, Last Visited April 2008.
minor modification. Ordinary nodes will have two states: sleep                         [4] M. Handley, V. Jacobson. SDP: Session Description Protocol, RFC
and active. After a long idle time an ordinary node will go to                            2327, Internet Engineering Task Force,April 1998. Available at:
                                                                                          http://www.ietf.org/rfc/rfc2327.txt
sleep mode by giving a sleep indication to its super node.                            [5] H. Schulzrinne, S. Casner, R. Frederick, V. Jacobson. RTP: A Transport
                                                                                          Protocol for Real-Time Applications, RFC 1889, Internet Engineering
C. Node Look Up                                                                           Task Force, January 1996. Available at: http://www.ietf.org/rfc/rfc1889.txt
                                                                                      [6] D. Bryan, B. Lowekamp. Decentralizing SIP, ACM Portal, Vol.5, Issue
   We can use two-tier caching scheme [24] to locate the target                           2, pages: 34-41, March 2007
                                                                                      [7] I. Stoica, R. Morris, D. Liben-Nowell, D.R. Karger, M.F. Kaashoek, F.
node and along with utilizing the parallel look up feature.                               Dabek, and H. Balakrishnan, Chord: A Scalable Peer-to-Peer Lookup Pro-
However, if the target node information is not cached in either                           tocol for Internet Applications, IEEE/ACM Transactions on Networking,
the cache of super node and ordinary node, then the super                                 page 17, 2003.
node will search using the parallel lookup feature of Bamboo                          [8] S. Rhea, D. Geels, T. Roscoe, and J. Kubiatowicz Kubiatowicz, Handling
                                                                                          Churn in a DHT In Proc. of USENIX Annual Technical Conference,
in order to locate the node. That will take less time to find the                          2004.
target node’s responsible super node.                                                 [9] 3. Sean Rhea, Byung-Gon Chun, John Kubiatowicz, and Scott Shenker,
                                                                                          Fixing the Embarrassing Slowness of OpenDHT on PlanetLab, Proceed-
                                                                                          ings of USENIX WORLDS 2005, December 2005
D. NAT Traversal Problem                                                              [10] A. Rowstron and P. Druschel. Pastry: Scalable, Distributed Object
                                                                                          Location and Routing for Large-Scale Peer-to-Peer Systems, In IFIP/ACM
  Super nodes are the most powerful nodes in terms of                                     International Conference on Distributed Systems Platforms, 2001.
available bandwidth, CPU processing power and they have a                             [11] B.Y. Zhao, L. Huang, J. Stribling, S.C. Rhea, A.D. Joseph, and J.D.
public IP address. In order to solve NAT traversal problem, a                             Kubiatowicz. Tapestry: A Resilient Global-scale Overlay for Service
                                                                                          Deployment, Selected Areas in Communications, IEEE Journal on, 2004.
super node can act as a relay for an ordinary node to establish                       [12] J. Hautakorpi, G. Camarillo. Evaluation of DHTs from the viewpoint of
session.                                                                                  interpersonal communications, ACM International Conference Proceed-
                                                                                          ing Series; Vol. 284, Proceedings of the 6th international conference on
                                                                                          Mobile and ubiquitous multimedia, pages 74-83, 2007
E. Connectivity Problem                                                               [13] D. Bryan, P. Matthews, E. Shim, and D. Willis. Concepts and Termi-
   Super nodes will use Bamboo to maintain the inter domain                               nology for Peer to Peer SIP draft-ietf-p2psip-concepts-01, Nov. 2007.
                                                                                          Expires: May 18, 2008.
overlay with other super nodes and regional algorithm like                            [14] D. Bryan, E. Shim, and B. Lowekamp. Use Cases for Peer-to-Peer
Chord, Pastry, etc. running in the inner domain overlay. Thus,                            Session Initiation Protocol (P2P SIP), draft-bryan-p2psip-usecases-00.txt
each Super node will maintain two DHT tables based on two                                 , July 2007. Expires: January 3, 2008
                                                                                      [15] D. Bryan, S. Baset, M. Matuszewski, and H. Sinnreich. P2PSIP Protocol
different algorithms. As a result the different domains running                           Framework and Requirements, draft-bryan-p2psip-requirements-00.txt ,
different DHT algorithm can communicate.                                                  June 2007. Expires: Jan 2008.




                                                                                 98                                   http://sites.google.com/site/ijcsis/
                                                                                                                      ISSN 1947-5500
                                                                 (IJCSIS) International Journal of Computer Science and Information Security,
                                                                 Vol. 9, No. 6, June 2011


[16] K. Singh and H. Schulzrinne. Peer-to-peer Internet Telephony Using SIP,
    In NOSSDAV ’05: Proceedings of the international workshop on Network
    and operating systems support for digital audio and video, pages 63-68.
    ACM Press, June 2005.
[17] K. Singh and H. Schulzrinne. SIPpeer: a session initiation protocol
    (SIP)-based peer-to-peer Internet telephony client adaptor, White paper,
    Computer Science Department, Columbia University, New York, NY, Jan
    2005. http://www.cs.columbia.edu/ kns10/publication/sip-p2pdesign. pdf.
[18] J. Rosenberg. Interactive Connectivity Establishment (ICE): A Method-
    ology for Network Address Translator (NAT) Traversal for Offer/Answer
    Protocols, draft-ietf-mmusic-ice-08. 2006.
[19] K. Singh and H. Schulzrinne. Peer-to-peer Internet telephony using
    SIP, Technical Report CUCS-044-04, Department of Computer Science,
    Columbia University, New York, NY, Oct. 2004.
[20] D. A. Bryan, B. B. Lowekamp, and C. Jennings. SOSIMPLE: A Server-
    less, Standards-based, P2P SIP Communication System In Proceedings
    of the 2005 International Workshop on Advanced Architectures and
    Algorithms for Internet Delivery and Applications (AAA-IDEA 2005),
    June 2005.
[21] L. Le, G. Kuo. Hierarchical and Breathing Peer-to-Peer SIP Sys-
    tem Communications, 2007, ICC ’07, IEEE International Conference,
    Pages:1887 - 1892, June 2007.
[22] K.K.Dhara, V. Krishnaswamy, S. Baset. Dynamic peer-to-peer overlays
    for voice systems, Pervasive Computing and Communications Workshops,
    2006. PerCom Workshops 2006. Fourth Annual IEEE International Con-
    ference, March 2006.
[23] Public      Key      Infrastructure(PKI)     tutorial.   available     at:
    http://www.cs.gmu.edu/ hfoxwell/EC511/pki.pdf. Last Visited- April
    2008.
[24] S. Pack, K. Park, T. Kwon, Y. Choi. SAMP: scalable application-
    layer mobility protocol, IEEE Communications Magazine, Vol. 44, Issue
    6,Page(s):86 - 92. JUNE 2006.
[25] J. Shi; Y. Wang; L. Gu; L. Li; W. Lin; Y. Li; Y. Ji; P. Zhang.
    A Hierarchical Peer-to-Peer SIP System for Heterogeneous Overlays
    Interworking, IEEE Global Telecommunications Conference. Page(s):93
    -97, November 2007.
[26] H. Jie, H. Yongfeng, L. Xing. MSPnet: Manageable SIP P2P media
    distribution system, Journal of electronics(China). Volume 24, November,
    2007
[27] A. Fessi, H. Niedermayer, H. Kinkelin, G. Carle. A Cooperative SIP In-
    frastructure for Highly Reliable Telecommunication Services, IPTCOMM,
    Proceedings of the 1st international conference on Principles, systems and
    applications of IP telecommunications, July 2007.




                                                                                  99                        http://sites.google.com/site/ijcsis/
                                                                                                            ISSN 1947-5500