Flexible bloom for searching textual content

Available Online at www.ijcsmc.com
International Journal of Computer Science and Mobile Computing
A Monthly Journal of Computer Science and Information Technology
ISSN 2320–088X
IJCSMC, Vol. 2, Issue. 5, May 2013, pg.208 – 213
RESEARCH ARTICLE
© 2013, IJCSMC All Rights Reserved 208
Flexible Bloom for Searching Textual Content
Based Retrieval System in an Unstructured P2P
Overlay Networks
S. Uvaraj1
, S. Suresh2
, Dr. E. Mohan3
1
M.E/CSE, Arulmigu Meenakshi Amman College of Engineering, Kanchipuram, India
2
B.Tech/IT, Sri Venkateswara College of Engineering, Chennai, India
3
Prof/CSE, Pallavan College of Engineering, Kanchipuram, India
Abstract— By using the fullest of a hybrid P2P protocol, Bloom Cast makes copies of the contents in the
network uniformly at a random across the P2P networks in order to achieve a guaranteed recall at a
communication cost of the network. BloomCast model works only when the two constraints are met: 1) the
query replicas and document replicas are randomly and uniformly distributed across the P2P network; and
2) every peer knows N, the size of the network. To support random node sampling and network size
estimation, BloomCast mixes a lightweight DHT into an unstructured P2P network. Further to reduce the
replication cost, BloomCast utilizes Bloom Filters to encode the entire document. Bloom Cast hybridizes a
lightweight DHT with an unstructured P2P overlay to support random node sampling and network size
estimation. Since P2P networks are self-configuring networks with minimal or no central control, P2P
networks are more vulnerable to malwares, malicious code, viruses, etc., than the traditional client-server
networks, due to their lack of structure and unmanaged nature. All peers in a P2P network is identified by its
identity certificates. The identity here is attached to the repudiation of a given peer. Self-certification helps us
to generate the identity certificate, thus here all the peers maintain their own and hence trusted certificate
authority which issues the identity certificate to the peer.
Key Terms: - Bloom Cast; Bloom Filters; Self-Certification; Self-Configuring Networks; Unstructured P2P
network
I. INTRODUCTION
A P2Pnetwork has also shown great potential to become a popular network tool for sharing information on
the Internet. Existing P2P full-text search schemes can be divided into two types: DHT based global index and
federated search engine over unstructured protocols. DHT-based searching engines are based on distributed
indexes that partition a logically global inverted index in a physically distributed manner. Due to the exact
match problem of DHTs, such schemes provide poor full-text search capacity. In federated search engines over
unstructured P2Ps, queries are processed based on flooding. The best candidate for supporting full-text retrieval
because the query evaluation operations can be handled at the nodes that store the relevant documents. Recall is
not guaranteed with acceptable communication cost using a flooding-based scheme.
II. RELATED WORKS
The content retrieval scheme is an important issue in the distributed P2P information sharing systems. There
are two content searching schemes in the existing P2P systems. For a structured P2P networks it is DHT-based
distributed global inverted index, for unstructured P2P networks we use federated search engines.

S. Uvaraj et al, International Journal of Computer Science and Mobile Computing Vol.2 Issue. 5, May- 2013, pg. 208-213
2.1. Pure P2P Systems
A pure P2P network does not have the notion of clients or servers but only equal peer nodes that
simultaneously function as both "clients" and "servers" to the other nodes on the network. The network
arrangement of this model differs from the client–server model. Here the communication is from and to the
central server. File Transfer Protocol (FTP) is a typical example of a file transfer that does not use the P2P
model. Here the client and server programs are distinct: the clients initiate the transfer, and the servers fulfil
these requests. The P2P overlay network consists of all the participating peers as network nodes. If there exists a
link between any two nodes that know each other: i.e. if a peer knows the location of another peer in a P2P
network, then there forms a directed edge between the former node and the latter in the overlay network. Based
on the linking between the various nodes in the overlay network, we can classify the P2P networks as structured
or unstructured.
2.2. Searching in Structured Networks
Structured P2P networks employ a globally consistent protocol to ensure that any node can efficiently route a
search to some peer that has the desired resource or data, even if it a rare one. But this process needs more
structured pattern overlay links. The most commonly seen structured P2P network is to implement a distributed
hash table (DHT), in which deviating of hashing is used to assign ownership of files to that particular peer. It is
not similar to the traditional hash table assignment in which in a for a particular array slots a separate key is
assigned. The term DHT is generally used to refer the structured overlay, but DHT is a data structure that is
implemented on top of a structured overlay.
2.3. Searching in Unstructured Networks
Unstructured P2P networks are formed when the overlay links are established randomly. The networks here
can be easily constructed by copying existing links of another node and then form its own links over a time. In
an unstructured P2P network, if a peer wants to find out a desired data in the network, the query is flooded
through the network which finds many peers that share their data. The major disadvantage here is that the
queries may not be resolved frequently. If there exists popular content then the available peers and any peer
searching for it is likely to find the same thing. In cases where a peer is looking for rare data shared by only a
few peers, then it is highly improbable that search will be successful. Since the peer and the content
management are independent of each other, there is no assurance that flooding will find a peer that has the
desired data. Flooding causes a high amount of signalling traffic in the network. These networks typically have
very poor Content Based Retrieval in Unstructured P2P Overlay Networks www.theijes.com The IJES Page 320
Search efficiency. Most of the popular P2P networks are unstructured.
2.4. Indexing and Source Discovery
The older P2P networks replicate the resources across each node in the network that is configured to carry out
the type of information. This will provides the local searching, but it requires much more traffic. Nowadays, the
modern networks using the central coordinating servers and it provide the search requests directly. Central
servers are mainly used for list out the potential peers that are present in the network, organizing their activities,
and searching purposes. In decentralized type of networks, searching was first done by flooding the search
requests along the peers. But now the new and more efficient search strategies called super nodes and
distributed hash tables are used.
2.5. Overlay Networks
An overlay network is a type of the computer network which is built on the top of another existing network.
Nodes that are present in the overlay network can be thought of as being connected as virtual links or logical
links, each one of which is corresponds to an appropriate path, connected through many physical links, in the
existing network. The major applications of overlay networks are, distributed systems such as cloud computing,
peer-to-peer systems, and client-server systems, because they are run on top of the Internet. Initially the internet
was built as an overlay network upon the telephone network whereas nowadays with the invention of VoIP, the
telephone network is turning into an overlay network that is built on top of the Internet. The area in which the
overlay networks used is telecommunication and internet applications.
III. SYSTEM APPROACH
Types of Nodes is an interactive system which provides detect text file from copyright infringement in P2P
file sharing by using bloomcast scheme.
Bootstrap node maintains a local repository and maintains the partial list of bloom cast nodes.
Normal peers to provide services of random node sampling and network size estimation.
Good connectivity and long uptimes are promoted to structured peers by bootstrap peers to forms a global
DHT.

3.1 Node Creation
Peer-to-peer (P2P) computing or networking is a distributed application architecture that divides the tasks
among the peers. Peers are active and more privileged participants in the application. They are said to form a
P2P network of nodes. These P2P applications become popular due to some files sharing systems such as
Napster. This concept paved a way to new structures and philosophies in many areas of human interaction. Peer-
to-peer networking has no restriction towards technology. It covers only social processes where peer-to-peer is
dynamic. In such context peer-to-peer processes are currently emerging throughout society. Peer-to-peer
systems implement an abstract overlay network which is built at Application Layer on the top of the physical
network topology. These overlays are independent from the physical network topology and are used for
indexing and peer discovery. The contents are shared through the Internet Protocol (IP) network. Anonymous
peer-to-peer systems are interruption in the network, and implement extra routing layers to obscure the identity
of the source or destination of queries.
3.2 Bloomcast
Bloom Cast is a novel replication strategy to support efficient and effective full-text retrieval. Different from
the WP scheme, random node sampling of a lightweight DHT is utilized by the Bloom Cast. Here we generate
the optimal number of replicas of the content in the required workspace. The size of the networks is not
depending on any factor since it is an unstructured P2P network. The size of the network is represent here as N.
By further replicating the optimal number of Bloom Filters instead of the raw documents, Bloom Cast achieves
guaranteed recall rate which results in reduction of the communication cost for replicating. We can design a
query evaluation language to support full-text multi keyword search, based on the Bloom Filter membership
verification. Bloom Cast hybrid P2P network has three types of nodes: they are structured peers, normal peers,
and bootstrap peers. A Bloom Cast peer stores a collection of documents and maintains a local storage also
known as repository. A bootstrap node maintains a partial list of Bloom Cast nodes it believes are currently in
the system. In previous P2P designs, there are different ways to implement the bootstrap mechanism. Bloom
Cast is an inter-domain protocol, operating between border routers in the AS hosting the source (source AS) and
the border routers of the ASes hosting the receivers (receiver ASes). However, for the sake of simplicity, we
will treat each AS a single node.
Algorithm 1 Query Evaluation of BloomCast
1: R ←Ø;
2: for all BFs replicated in this peer do
3: BooleanContainFlag ← True
4: for all terms in Q do
5: if ∃(j)(1 ≤ j ≤ k)s.t.BFx[hj(t)] = 0 then
6: ContainFlag ← False;
7: end if
8: end for
9; if ContainFlag True then
10: R ← R U {urlx};
11: end if
12: end for
13: return R
3.3 Stemming Algorithm
Stemming is the process for reducing inflected words to their stem, base or root form generally a written
words form. Many search engines treat words with the same stem as synonyms as a kind of query broadening a
process called conflation.
Function: Stemming is a process of reducing a word by removing some pattern. For example : when user
searches with keyword 'Searching' then the stemming process will remove the 'ing' from 'searching' and you will
get the 'search'. Then you can use this keyword 'search' to use for searching in the index server. It’s done using
porter algorithm.
Input: A query with collection of keywords.
Output: keywords are stemmed to their roots and used for the search system.

3.4 Bloom Filter
Bloom Filters to encode the transferred lists while recursively intersecting the matching document set. A
Bloom Filter is an efficient data structure method that is used to test whether the element belongs to that set or
not. False positive retrieval results are also possible, but false negatives are not possible; i.e. a query returns
either it is „inside the set‟ or „not inside the set‟. Elements can only be added to the set and cannot be removed.
When more elements are added to the set then the probability of false positives increases. Bloom Casting is a
secure source specific multicast technique, which transfers the membership control and per group forwarding
state from the multicast routers to the source. It uses in-packet Bloom filter (iBF) to encode the forwarding tree.
Bloom Casting separates multicast group management and multicast forwarding.
It sends a Bloom Cast Join (BC JOIN) message towards the source AS. The message contains an initially
empty collector Bloom filter. While the message travels upstream towards the source, each AS records
forwarding information in the control packet by inserting the corresponding link mask into a collector. After this,
it performs a bit permutation on the collector. The figure for Bloom Filter and their memory storage is designed
here to show the interconnections between source and specific multicast protocols. Unlike traditional IP
multicast approaches, where the forwarding information is installed in routers on the delivery tree, in Bloom
Cast, transit routers do not keep any group-specific state.
Fig. 1. Bloom Filter
Efficient probabilistic data structure that is used to test whether an element is a member of a set.
False positive retrieval results are possible.
False negative are not.
Algorithm 2 Cast Bloom Filters
Require: EstimatedNetworkSize N is achieved
1: for all documents in local collection do
2: create an empty bit vector with m bits for document x,BF,
3: for all terms in a document do
4:insert term t into BF, by setting the hj (t)th bits of BF, to 1, where {hj (.), 1 < j < k} is the set of hash
functions used by BF,
5: end for
6: end for
7: sample an optimal number of r, random
peers in the network by the lightweight DHT;
8: replicate BF, together with url, the URL of document x, to the set of randomly sampled nodes;
9: return
3.5 Distribute BF among the Nodes
Once the data are converted into the URL’s, the url’s are distributed to all other nodes. Once the node the
request for the particular data in the network, once search has been finished, the best results will display to the
user.

3.6 Ranking Process
The Ranking of data, so that the new users may able to find the exact data when they search/surfing. Using
the chord algorithm, the peer node will do forward and backward search and as a result each document is
provided with the rank and hence according to the rank given, the best document is identified by the server.
3.6.1 Retrieval of Data
By analyzing the results of the bloom hash table, we found that this is significantly faster than a normal hash
table using the same amount of memory.
IV. DATAFLOW DIAGRAM
Fig. 2. Dataflow Diagram
V. CONCLUSION AND FUTURE ENHANCEMENT
We here propose an efficient and effective full-text retrieval scheme in an unstructured P2P networks using
Bloom Cast method. Bloom Cast is effective here because it guarantees the recall with high probability. The
overall communication cost of a full-text search is reduced below a formal bound. Thus it is efficient and
effective among other schemes. Furthermore the communication cost for replication is also reduced since we
replicate Bloom Filters instead of the raw documents across the network. We demonstrate the power of Bloom
Cast design through both mathematical proof and comprehensive simulations based on the TREC WT10G data
collection and query logs from a real world search engine. Peer-to-peer (P2P) networks are self-configuring
networks with minimal control. P2P networks are more vulnerable to dissemination of malicious code, viruses,
worms, and Trojans than the traditional client-server networks, due to their unregulated and unmanaged nature.
All peers in the P2P network are identified by their identity certificates i.e. aka identity. The reputation of a
given peer is attached to its identity. The identity certificates are generated using self-certification, and all peers
maintain their own (and hence trusted) certificate authority which issues the identity certificate(s) to the peer.

VI.RESULTS
To evaluate the performance of BloomCast, in the simulation implement three baseline schemes.
Fig. 3. Recall
Fig. 4. Query traffic
Fig. 5. Efficiency
The result in Fig. 4 shows that the average query traffic of Bloom Cast is 6:5 _ 105, very similar with that of
the WP algorithm. The average traffic of BloomCast is much less than that of flooding.
REFERENCES
[1] E. Cohen and S. Shenker, “Replication Strategies in Unstructured Peer-to-Peer Networks,” Proc. ACM
SIGCOMM ’02. pp. 177-190, 2002.
[2] H. Shen, Y. Shu, and B. Yu, “Efficient Semantic-Based Content Search in P2P Network,” IEEE Trans.
Knowledge and Data Eng., vol. 16, no. 7, pp. 813-826, July 2004
[3] R.A. Ferreira, M.K. Ramanathan, A. Awan, A. Grama, and S.Jagannathan, “Search with Probabilistic
Guarantees in Unstructured Peer-to-Peer Networks,” Proc. IEEE Fifth Int’l Conf. Peer to Peer
Computing (P2P ’05), pp. 165-172, 2005.
[4] S. Robertson, “Understanding Inverse Document Frequency: On Theoretical Arguments for IDF,” J.
Documentation, vol. 60, pp. 503- 520, 2004.
[5] P. Reynolds and A. Vahdat, “Efficient Peer-to-Peer Keyword Searching,” Proc. ACM/IFIP/USENIX
2003 Int’l Conf. Middleware (Middleware ’03), pp. 21-40, 2003.
[6] D. Li, J. Cao, X. Lu, and K. Chen, “Efficient Range Query Processing in Peer-to-Peer Systems,” IEEE
Trans. Knowledge and Data Eng., vol. 21, no. 1, pp. 78-91, Jan. 2008.
[7] I. Stoica, R. Morris, D. Karger, M.F. Kaashoek, and H.Balakrishnan, “Chord: A Scalable peer-to-Peer
Lookup Service for Internet Applications,” Proc. ACM SIGCOMM ’01, pp. 149-160, 2001.
[8] J.P.C. Jie Lu, “Content-Based Retrieval in Hybrid Peer-to-Peer Networks,” Proc. 12th Int’l Conf.
Information and Knowledge Management (CIKM), pp. 199-206, 2003.
[9] E.M. Voorhees, “Overview of Trec-2009,” Proc. 16th Text Retrieval Conf. (TREC-11), 2009.
[10] A. Broder and M. Mitzenmacher, “Network Applications of Bloom Filters: A Survey,” Internet Math.,
vol. 1, no. 4, pp. 485-509, 2004.

Flexible bloom for searching textual content

More Related Content

What's hot (15)

Viewers also liked (14)

Similar to Flexible bloom for searching textual content (20)

Recently uploaded (20)

Flexible bloom for searching textual content