This document summarizes and compares four research papers related to Domain Name System (DNS) traffic and performance. The original 1991 research paper analyzed DNS traffic captured from root name servers and found that DNS consumed 20 times more bandwidth than necessary due to errors. It classified seven types of DNS implementation errors. The later papers proposed techniques like enhanced round robin DNS load balancing, a CDN-based DNS architecture, and a federated name resolution framework to address DNS issues in different network architectures and improve performance.

1. Report on
An Analysis of wide-area Name Server traffic
Subject
Advance Operating System
Submitted To
Mr. Asim Munir
Class
MSCS-F14
Prepared By:
Sunawar Khan
Reg. No: 813-MSCS-F14
Date
27th May, 2015

2. Table of Contents
Abstract...........................................................................................................................................3
Introduction....................................................................................................................................3
Literature Survey...........................................................................................................................3
Research Paper #1:....................................................................................................................4
Research Paper #2:....................................................................................................................4
Research Paper #3:....................................................................................................................5
A Comparative Study: ...................................................................................................................6
References:.....................................................................................................................................9

3. Abstract
World largest distributed database in DNS, millions of computers are connected with
each other. DNS is fairly large source of wide-area RPC-Like Traffic.
In year 1991,a research paper reported that eight percent of the packets and four
percent of the bytes are travel across NFS net due to DNS. In this paper, the author
formally discussed the performance of DNS, caching, retransmission timeout
calculation. Using different algorithms, how can DNS resiliency lead to disastrous
consequences when certain implementation faults are trigged?
Introduction
DNS associates hostnames with their respective IP addresses, so as when users want
to connect to other machines on the network, they can refer to them by name without
having to remember IP addresses.
The Domain Name System (DNS) is the world’s most distributed database. DNS is a
distributed, replicated name service whose primary purpose is to map host names into
corresponding Internet addresses, map Internet addresses into hostnames. Real Word
Problem related to DNS is that, DNS consumed twenty times more wide-area network
bandwidth unnecessarily.
DNS client software is known as a resolver. At least a dozen resolver implementations
exist. Some of these implementations have been re-released to repair various bugs.
Some of these implementations have been re-released to repair various bugs.
Caching: DNS server cache responses of successful quires.
Retransmission Algorithms: All DNS transactions employ the UDP unreliable datagram
protocol to avoid the additional messages and latency to set up and tear down a TCP
connection.
Resolver algorithm: The resolver sends the query to the first of the three servers and
waits for a timeout.
Name Server Algorithm: A canonical name is an alias for the query name; for example,
castor. USC.edu is an alias for girtab.USC.edu. The server limits itself to a total of 20
referrals or 8 canonical name substitutions.
LiteratureSurvey
As a part of literature review we studied several Domain Name Service related papers
and have chosen three latest papers. The motivation behind right selection of papers is

4. to identify the relationship among first paper (published in 1991) till up-to-date research
being advanced in which manner and direction.
Research Paper #1:
Research paper [2]
addresses about an enhancement to the conventional Round Robin
DNS load balancing technique which allows a single domain name to be associated
with several web server IP addresses in a rotated order in a server cluster. Once the
DNS server resolves the domain name to one of the web server IP address, the
subsequent requests from the same client will be sent to the same server regardless of
the current condition of the server. However, this conventional load balancing technique
has a few disadvantages.
Approach: The paper presents a new approach to enhance the DNS load balancing
services to be more intelligent in load distribution. The DNS server is taking
consideration of the status of the servers and will be distributing the services requests
based on the performance matrix of the servers in the cluster.
Technique enhanced: The Round Robin DNS Load Balancing is a technique that
providing a high level of availability of some services such as webs site, ftp or others
services. It also uses to balance the load between the multiple servers to prevent
overloading a single server. In this paper a new approach is proposed to enhance the
Round Robin DNS load balancing by introducing a new performance indexing algorithm
and a task distribution scheduler at the DNS server.
Results are presented in controlled environment using Simulation.
Conclusion: performance of a load balancing system can be further enhanced if the
right parameters are being taken into account in task or request distribution. Hence, the
proposed algorithm in paper has indicated that by optimizing the load balancing scheme
based on the right performance parameters, higher system efficiency can be achieved.
Research Paper #2:
Research paper [3]
enhances end-user experiences, Content Distribution Networks
(CDNs) effectively exploit the Domain Name System (DNS) to redirect end-users to
close-by content replicas over short time scales. While the use of DNS has brought a
significant advantage to CDNs, in this paper we confirm that reliance on DNS also
poses a fundamental threat to large-scale CDNs’ content distribution model. In
particular, we demonstrate that a considerable penetration of public DNS resolving
services (e.g., OpenDNS and GoogleDNS)effectively corrupts the CDN approach,
equally the large-scale server distribution and quick DNS redirections.

5. Technique enhanced in this paper: First, we systematically evaluate and quantify how
the use of publicDNS resolving services impacts Akamai’s content distribution model,
the corresponding end-user service performance, and ISPs that host CDN edge
servers. Second, we show that a CDN-based DNS architecture, which can effectively
function as both the current authoritative name servers and resolvers, coexist with the
current DNS infrastructure, and directly response end-users with authoritative DNS
records. Third, we implement a prototype CDN-based DNS system by using popular
Web 2.0 websites as a vehicle to publish DNS records onto CDN edge servers. We
demonstrate that in an extreme setting, such an approach can achieve one order of
magnitude faster lookup times relative to existing DNS resolving systems.
New Approach Proposed: This paper demonstrates that CDNs are not only to be
another DNS provider, but also can comprehensively address DNS problems and
provide a fundamentally different DNS service, yet superior. Some of the CDNs (e.g.,
Google) have involved the name resolving service and they have strong incentive to
improve their services.
In this paper a comparison of DNS architectures is performed and a new Prototype
implementation is proposed and measured in terms of Update latency, Update traffic for
the domain management and Availability.
Results are presented in the comparison table using Percentage of changed DNS data
per day.
Conclusion: In this paper, we demonstrated that the increasing use of public DNS
systems, such as GoogleDNS and OpenDNS, fundamentally disrupts CDNs’ content
distribution model. While we have confirmed that such systems indeed improve the
DNS lookup times, and hence are beneficial for the vast majority of websites, we
showed that this is not the case for the websites served on CDNs. Indeed, because
end-users get redirected to a small number of heavily sub-optimal replicas closest to
public DNS resolvers, not the end-users, significant problems arise.
Research Paper #3:
Research paper [4] addresses about a key problem in all name resolution protocols
today is that no one protocol performs well across all network architectures. In addition,
DNS, the most widespread solution today, depends on a static and connected network
layer and faces significant challenges in dynamic wireless networks.

6. Technique enhanced: This paper introduces FERN(Federated Extensible Resolution of
Names), the first framework designed to enable efficient name resolution across
heterogeneous systems operating in dynamic or static networks. FERN organizes
nodes into name resolution groups and allows each group to perform name resolution
independently and in a manner best suited for that group. FERN arranges these name
resolution groups into a hierarchy and allows these groups to communicate efficiently,
discover each other’s presence, and resolve each other’s names.
We demonstrate the flexibility and interoperability of FERN by deploying and evaluating
it across heterogeneous environments, including a MANET, an infrastructure-based
wireless network, and the Internet.
Contribution: Paper shows that FERN performs at least as well as DNS, and yet
extends name resolution to networks in which DNS is inadequate. This paper presents
FERN NRG rules and then builds FERN name resolution tree. Most importantly, it
proofs correctness of proposed model.
Conclusion: FERN is novel in its ability to interface radically different name resolution
architectures. By providing a unifying framework for these protocols, we have laid a
foundation for interoperability between future name resolution protocols that are highly
specialized for a particular network environment
A Comparative Study:
A paper that was published in 2012 tells us, Content Distribution Networks (CDNs)
effectively exploit the Domain Name System (DNS) to redirect end-users to close-by
content. Proposed solution in this paper is, quick DNS redirections. Systematically
evaluate and quantify how the use of public DNS resolving services impacts Akamai’s
content distribution model, the corresponding end-user service performance, and ISPs
that host CDN edge servers. CDN-based DNS architecture, which can effectively
function as both the current authoritative name servers and resolvers, coexist with the
current DNS infrastructure. Paper demonstrates that increasing use of public DNS
systems, such as GoogleDNS and OpenDNS, fundamentally disrupts CDNs’ content
distribution model. Indeed, because end-users get redirected to a small number of
heavily sub-optimal replicas closest to public DNS resolvers, not the end-users,
significant problems arise.
Another Article tells us that, there are different problem in name resolution protocol
performs well across all network architectures. FERN arranges these name resolution
groups into a hierarchy and allows these groups to communicate efficiently, discover
each other’s presence, and resolve each other’s names. FERN is novel in its ability to
interface radically different name resolution architectures.

7. It just provides a unifying framework for these protocols and lays a foundation for
interoperability between future name resolution protocols.
Another technique says that Round Robin DNS load balancing technique which
allows a single domain name to be associated with several webserver IP addresses in a
rotated order in a server cluster.
The performance of a load balancing system can be further enhanced if the right
parameters are being taken into account in task or request distribution. Hence, the
proposed algorithm in paper has indicated that by optimizing the load balancing scheme
based on the right performance parameters, higher system efficiency can be achieved.
The Research Paper (original):
Research paper [1]
is a study is based on two 24-hour packet traces collected from DNS
root name server a.isi.eduand three other name servers that replicated various
domains. Trace Collection - collection machine used a network interface tap (NIT)
program to collect all TCP and UDP DNS packets, including non-transit local traffic.
End to End Loss Rate - method to calculate end-to-end loss rate because our collection
machine’s network interface does not count dropped DNS packets. Although the end to-
end loss rate that we calculated was below 1%.Performance - DNS consumes at least
twenty times more wide-area bandwidth than is strictly necessary. Classifying DNS
errors, presents the classification scheme that devised, and discusses wide-area DNS
performance using our scheme.
• Caching
• Retransmission Algorithm
• Resolver Algorithm
• Name Server Algorithm
• The Net Effect
How different bugs can classify? The name servers probably are not properly primed
with alternate routes, do not detect server failure, do not back off their retransmission
timeouts properly, probably don’t cache properly, and suffer from a form of looping
because they do not forward queries properly.
How we can reduce bandwidth? There are different types of errors that cause of high
usage of bandwidth. We can reduce these errors using different techniques. Seven
classes of DNS implementation errors that identified.
DNS implementation.

8. DNS Replication: DNS replication is the process of coping records from one DNS server
to another. Domain replicas do not run on individual servers, but that most servers
replicate several domains.
Conclusion: In proposed article wide-area network traffic due to DNS will decrease as
defective name servers and resolvers are replaced, assuming no vendor releases
another devastating bug. As old implementations are corrected. We are currently writing
such software for the Berkeley BIND name server. We hope to install this software into
a root name server in the coming months.

9. References:
1. An Analysis of Wide-Area Name Server Traffic by B.Danzing, Obraczka, A.
kumar, danzigtksc.edu, 1991.
2. Efficient Load Balancing for Bursty Demand in Web based Application Services
via DomainNameServices by Lu Chin, Chong Eng Tan, M. Imran Bandan,
University Malaysia Sarawak, 2010.
3. A CDN-based Domain Name System by Zhen Qinm Chunjing Xiao, Qiyaho
Wang, Computer Communications 45 (2014) 11–20, 2014
4. FERN: A unifying framework for name resolution across heterogeneous
architectures by spencer sevilla, priya Mahadevan, Computer Communications
56 (2015) 14–24, 2015.