Ijirsm bhargavi-ka-robust-distributed-security-using-stateful-csg-based-distributed-firewall

ISSN: XXXX-XXXX Volume X, Issue X, Month Year
A Robust Distributed Security Using Stateful CSG
Based Distributed Firewall
Bhargavi K
Dept of Computer Science and Engineering
BTL Institute of Technology
Bangalore, India
bhargu.116@gmail.com
Abstract
Distributed firewalls have been developed to maintain the
networks with a higher level of protection than conventional
firewalling mechanisms like gateway and host-based fire-
walls. even though distributed firewalls provide higher secu-
rity, they too have boundaries. Toovercome those limitations
we presenting the design & implementation of a new distrib-
uted firewall model, based on stateful Cluster Security
Gateway (CSG) architecture[1]. This distributed security
model are implemented in bottom-up approach means each
cluster of end-user hosts are secured first using the CSG
architecture. These different CSGs are then centrally man-
aged by the Network Administrator. A file-based firewall
update mechanism is used for dynamic real- time security.
IPsec protocol is used to secure the firewall policy update
distribution while X.509 certificates cater for sender/receiver
authentication. The major advantage of this approach is to
distributed security include tamper resistance, anti-spoofing,
anti-sniffing, secure up to date firewall updating, low overall
network load, high scalability and low firewall junction
times.
Keywords-stateful CSG architecture, distributed firewall,
distributed cluster security, Layer 2 per-packet load balancing
Introduction
Distributed firewalls have been developed to main-
tain the networks with a higher level of protection than con-
ventional firewalling mechanisms like gateway and host-
based firewalls. Although distributed firewalls are capable of
achieve their purpose, but they too have some limitations.
Problems like raise in processing load on end-user hosts
because of the packet filtering strain, decline in overall net-
work performance because of dynamic firewall updating,
and user tampering in particular, hinder the deployment and
usage of distributed firewalling solutions. To overcome the-
se limitations of distributed firewall we use a new approach
to distributed firewalling, based on the stateful CSG archi-
tecture. The CSG architecture [1] provides a technique for
grouping multiple networking elements together such as
routers, security gateways, and switches in order to create
more secure, more trustworthy switched network clusters.
The idea behind the CSG-based distributed firewall design is
that is robust security.
In this paper, is provided at the very cluster level, the whole
of the network will become more secure as we can reduce
the occurrence of both insider & external attacks, and limit
their spread & effects more readily. A 2-active-node stateful
CSG is used for protecting each end-user cluster in our
working prototype.
Distributed Firewall:
Pioneered by Steven Bellovin [2] in 1999, distrib-
uted firewalls have been created to overcome the drawbacks
of both gateway & host-based firewalls, and more specifical-
ly, in order to prevent insider attacks. According to Ioannidis
et al. [3], a distributed firewall is a mechanism that defines
centralized security policy but the later is applied at the edg-
es. Distributed firewalls are basically centrally managed
host- resident security software application that protect a
network’s essential endpoints against not needed intrusions.
The conceptual design of distributed firewalls rests upon
three elements:
1. A general policy language that is used for defining securi-
ty policies that are distributed to the firewall endpoints form-
ing the distributed firewall. Examples of general policy lan-
guages include KeyNote [4] and Firmato [5].
2. Network-wide techniques for the distribution and function
of the security policy files to the distributed firewall end-
points.

International Journal of Innovatory research in Engineering and Technology - IJIRET
ISSN: XXXX-XXXX Volume X, Issue X, Month Year 22
3. IPsec: security protocol that provides network-level en-
cryption for the secure transmission of the security policy.
A. Major Strengths
• Centralized management
Security policies are formulated centrally and then
distributed to the all endpoints for enforcement. security
policies over the network and control over their deployment
is enhanced and maintained [3].
• Defense in depth
When used together with the gateway firewall, dis-
tributed firewalls contribute for multiple layers of defense
that an malicious node has to pierce through. It allows time
to other defense mechanisms to counter the threat effective-
ly, and thus, delay & prevent its spread in the network [6].
B. Major Limitations
• User tampering
According to Wei Li [7], this discribe the biggest
issue in distributed firewalls. Users can change host-based
firewall rules at will or totally remove the firewall, thereby
revealing those hosts to attacks. Hackers can attack both
internal & remote hosts.
• Decrease in network performance
The utilization of updates of real-time security poli-
cy will add substantial strain on the network with all the
traffic that is being generated by the distributed firewall. As
a result, the network becomes more vulnerable to DoS at-
tacks [7].
• Increase in host load
There is degradation in host performance. The host-
level packet filtering adds considerable load on hosts with
limited resources.
• host-resident components of distributed firewalls having
high reconfiguration time Since distributed firewalls allow
for dynamic updating of security policies, the bigger the size
of a network, the more time it takes to re-deploy security
policies.
Figure 1. The stateful CSG-based distributed firewall architecture.
Design Of The Stateful CSG-Based
Distributed Firewall
The stateful CSG-based distributed firewall is as
shown in Fig. 1. This document distributed firewalling archi-
tecture contains of four main components, namely the Net-
work Administrator machine, the Cluster Security Manager
(CSM), the stateful CSG & the CSG-based gateway firewall,
and several sub- systems like the Policy Repository, the Pol-
icy Distributor & the Policy Handler. In our test implemen-
tation, each end-user cluster is protected by a 2-active-node
stateful CSG. Each CSG-protected end-user cluster possess-
es a dedicated CSM, whose main job is to receive firewall
updates from the Network Administrator and forward them
to the CSG firewall nodes falling under its responsibility.
The workings of these different components are described in
greater detail below:
A. The Network Administrator Machine This machine is
used by the Network Administrator for managing the various
network components. It is from this computer that the Net-
work Administrator updates CSG firewall nodes. This ma-
chine contains two major components – the Policy Reposito-
ry and the Policy Distributor.
The Policy Repository: The Policy Repository is a
essential database where all the firewall scripts executed in
the network are stored. All the firewall updates are also
stored there. The Network Administrator can thus contact
the available firewall scripts in order to create new firewall
update files when the network is under attack. Store The
firewall scripts and update files in usable formats so that
they can be directly applied on the firewall nodes. All scripts
and updates assign to a particular cluster are stored together
for easy referencing.
The Policy Distributor: The Policy Distributor is
used by Network Administrator for sending firewall updates
to Cluster Security Managers (CSMs). The Policy Distribu-
tor creates end-to-end connections with the appropriate
CSMs. These connections are authenticated & encrypted for
secure transmission of firewall updates across the network.
This helps in preventing threats like man-in-the-middle at-
tack.Firewall updates are send to all firewall nodes via
CSMs because direct updating of firewall nodes will require
secure connections between the Policy Distributor and the
firewall nodes themselves. This will considerably increase
the processing damage on the firewall nodes as they will
then be acting as IPsec gateways. IPsec packet processing,
together with Ebtables & IPTables packet filtering, will con-
siderably decrease the efficiency and throughput of the fire-

wall nodes. Unicast transmission mode is chosen over mul-
ticast as not all the CSMs will need updating at a particular
point in time. There may be firewall updates that are meant
only for perticular cluster and not for the other clusters in the
network. If all the CSMs are made part of a single multicast
group, then all of others will have to accept the firewall up-
dates. This will cause their respective rule sets to increase in
size unnecessarily, and this will potentially affect firewall
performance.
B. The Cluster Security Manager (CSM) The CSM is the
first and foremost receiver of firewall updates from the Poli-
cy Distributor. It is the endpoint of the secure connections
established by the Policy Distributor. Each end-user cluster
has exactly one CSM. The CSM consists of a user-level pro-
cess that waits for firewall updates from the Policy Distribu-
tor and then send them to the stateful CSG firewall nodes
which as responsibility.
C. The Stateful CSG It contains multiple active firewall
nodes working in parallel to filter traffic travelling to and
from the end-user hosts of a perticular cluster. The CSG
architecture uses a different type of load estimation – the
Ebtables distributed sender-initiated MAC-based per-packet
load balancing (PPLB) scheme, where the load balancing is
done by the end-user nodes themselves. PPLB helps make
optimum usage of network links by allowing for equal dis-
tribution of traffic along those links. This Layer 2 PPLB
scheme has been developed primarily for a seamless combi-
nation in load balancing setups involving stealth firewalls, It
load balances network traffic onto MAC addresses rather
than IP addresses. However, it can be successfully used in
IP-based networks as well. This scheme is advantageous as
it prevents the creation of single points of failure by remov-
ing the need for a dedicated load balancer, and it integrates
well in already-in-place switched networks so that no major
re-design is required. The CSG architecture deployed for
each end-user cluster provides the following security mech-
anisms:
• Layer 2 and Layer 3 packet filtering using Ebtables and
IPTables respectively.
• Network Access Control (NAC) using MAC ACLs applied
on specific switch ports [1] to ensure that end- user hosts
communicate only via the firewall nodes.
Port security [8] so as to prevent source MAC address
spoofing.
Implementation Details
The high-level architecture of the stateful CSG
based distributed firewall is shown in different software
components as follows:
A. The Policy Distributor ( /etc/dfw/poldis_d.c ) is imple-
mented as a user-space program written in C. The program
takes as argument the full pathname of firewall update file
and the IP addresses of the appropriate CSMs. TCP is used
for distributing the firewall updates to the CSMs as it pro-
vides secure delivery of packets, which is crucial to the de-
livery of firewall updates. The Policy Distributor uses the
polides_handleupdate() function to read the specified fire-
wall update file, and the polides_sendingupdate() function
sends the update out to each specified IP address. Figure 2.
High-level implementation details of the stateful CSG-based
distributed firewall.
Figure 2. High-level implementation details of the stateful CSG-based
distributed firewall.
B. The Cluster Security Manager (CSM) The CSM is made
up of two user-space parts - a firewall update receiving part
and a firewall update sending part. Both parts make use of
TCP sockets. The receiving & sending parts are implement-
ed in the /etc/dfw/csm.c file. In the receiving part, the pack-
ets from the Policy Distributor are read by the clse-
ma_handleupdate() function. The firewall update file is re-
constructed on the CSM so as to ensure that the file is re-
ceived error-free and in its entirety. If an error occurs, it is
logged and the Network Administrator will be notified. In
the sending part, the clsema_ sendingupdate() function is
invoked, which then sends the firewall update out to each of
the CSG firewall nodes via unicast TCP connections. The IP
addresses of CSG firewall nodes are kept in a file (
/etc/dfw/fwl_list.txt ) that is created by the Network Admin-
istrator on the CSM. The clsema_sendingupdate() function
reads fwl_list.txt in order to find the IP addresses to which
the updates have to be sent. Both active and backup firewall
nodes are updated. Any communication errors between the
CSM and the firewall nodes are logged and the Network
Administrator is notified of them. For example, if ever one
or more of the firewall nodes becomes inaccessible to the
CSM when clsema_sendingupdate() tries to send a particular
firewall update, say because of damaged cable or NIC fail-
ure on the firewall node(s), the connection error thereby
generated is caught and recorded in a log file, which is even-
tually sent to the Network Administrator. Multicasting is not
used in the sending part because UDP does not provide reli-
able packet delivery. Reliability can, nevertheless, be added

to a UDP multicast application by applying features like
positive acknowledgements, lost packet retransmission, use
of sequence numbers and packet re-ordering. However, the
program complexity [9] is increased adding reliability,
which is not justified when it comes to updating such small
numbers of firewall nodes per cluster. The CSM is deployed
in a failover configuration such that if it fails, a backup ma-
chine will take up its place. Keepalived [10] is used for this
purpose. For our working prototype, a 2-active-node stateful
CSG is utilized to secure each end-user cluster as shown in
[11].
C. The Stateful CSG-based Gateway Firewall A 4-active-
node stateful CSG is used to implement the gateway fire-
wall. The main difference between the CSG used for protect-
ing an end-user cluster and the CSG for the gateway firewall
is that, in the latter implementation, the four active firewall
nodes are sandwiched between two routers which are re-
sponsible for load balancing network traffic on a per-packet
basis onto the firewall. Static routing, enabling CEF, and
using the "ip load-sharing per-packet" command on the
router interfaces help achieve per-packet round-robin load-
balancing.
D. IPsec
In our design, IPsec is used for securing the distribution of
firewall updates from the Network Administrator machine to
CSMs, and for securing error results from CSMs to the Net-
work Administrator. The paths between a CSM and its clus-
ter’s CSG firewall nodes are not IPsec-protected because
the firewall nodes accept firewall updates related for their
local process only from the related CSM of the cluster.it
only accepted from the eth2 interface on the firewallnodes.
Each CSM communicates only with Network Administrator
machine and the CSG firewall nodes falling under its con-
trol. The Network Administrator machine’s analyzed using
the utilization of digital certificates. Direct IPsec connec-
tions are not established with the firewall .This processing is
handled by the CSM. The latest Linux kernels provide native
support for IPsec. IPsec is installed and The cryptographic
algorithms in the CryptoAPI hav been applied in the IPsec.
Transport mode is used to for send the secure error log in-
formation from CSM to system Administrator. The IPsec-
tools package contains the IKE
daemon, racoon, and the setkey utility [12]. racoon is used
for the arranging up of automatically keyed IPsec connec-
tions .
Concluding Remarks
This paper describes about a new approach to dis-
tributed cluster and network security – the stateful CSG-
based distributed firewall architecture. This distributed secu-
rity model successfully overcome the limitations plaguing
distributed firewalls and apply few desirable characteristics
of its own. In this process we come across the Policy reposi-
tory and policy distributer witch is good in database man-
agement and administration respectively. Future work will
related to the implementation of load balancing network
interface cards (LB-NICs). In order to control the tampering
of the Ebtables load balancing rules that are available on the
end-user hosts by malicious users, the Ebtables MACbased
load balancing structure can be incorporated into a tamper-
resistant network interface card with on-board processing
engines. This approach adopts the implementation of hard-
ware-based distributed firewalls like EFW [13], [14] and
ADF [15].
References
[1] V. Ramsurrun, and K. M. S. Soyjaudah, “Efficient cluster security
gateway architecture for per-packet load balanced IP filtering on
switched clusters,” in 2006 Proc. CSNDSP’06 Conf., pp. 256-261.
[2] S. M. Bellovin, “Distributed firewalls,” ;login: magazine, special issue
on security, 1999.
[3] S. Ioannidis, A. D. Keromytis, S. M. Bellovin, and J. M. Smith,
“Implementing a distributed firewall,” in Proc. 7th ACM Conf.
Computer and communications security, Athens, 2000, pp. 190-199.
[4] M. Blaze, J. Feigenbaum, J. Ioannidis, and A. Keromytis, “The KeyNote
Trust-Management System Version 2,” RFC (Informational) 2704,
Internet Engineering Task Force, 1999.
[5] Y. Bartal, A. Mayer, K. Nissim, and A. Wool, “Firmato: A Novel
Firewall Management Toolkit,” in Proc. IEEE Symposium Security and
Privacy, 1999, pp. 17-31.
[6] M. Gangadharan, and K. Hwang, “Micro-firewalls for dynamic network
security with distributed intrusion detection,” in IEEE Int. Symposium
Network Computing and Applications (NCA'01), 2001.
[7] W. Li (2000). Distributed Firewall [Online]. Available:
http://citeseer.ist.psu.edu/li00distributed.html
[8] W. Odom, “CCENT/CCNA ICND1 Official Exam Certification Guide,”
2nd ed. Indianapolis, USA: Cisco Press, 2008, pp. 253-256.
[9] W. R. Stevens, B. Fenner, and A. M. Rudoff, Unix Network
Programming: The Sockets Networking API Volume 1, 3rd ed., Addison-
Wesley, 2004, pp. 595-598.
[10] Keepalived website (2007). Keepalived for Linux - Linux High
Availability [Online]. Available: http://www.keepalived.org/index.html
[11] V. Ramsurrun, and K. M. S. Soyjaudah Electrical & Electronic Engi-
neering Department University of Mauritius (UoM) Réduit, Mauritius
[12] R. Spenneberg (2007). IPsec HOWTO (Revision 0.9.96) [Online].
Available: http://www.ipsec-howto.org/
[13] T. Markham, L. Meredith, and C. Payne, “Distributed embedded
firewalls with virtual private groups,” in 2003 Proc. 3rd DARPA
Information Survivability Conf., vol. 2, pp. 81–83.
[14] C. Payne, and T. Markham, “Architecture and applications for a
distributed embedded firewall,” in Proc. 17th Annual Computer Security
Applications Conference (ACSAC 2001), 2001, pp. 329-336.
[15] L. M. Meredith, “A summary of the autonomic distributed firewalls
(ADF) project,” in 2003 Proc. 3rd DARPA Information Survivability
Conf., vol. 2, pp. 260–265.

Ijirsm bhargavi-ka-robust-distributed-security-using-stateful-csg-based-distributed-firewall

More Related Content

What's hot (16)

Viewers also liked (17)

Similar to Ijirsm bhargavi-ka-robust-distributed-security-using-stateful-csg-based-distributed-firewall (20)

More from IJIR JOURNALS IJIRUSA (6)

Recently uploaded (20)

Ijirsm bhargavi-ka-robust-distributed-security-using-stateful-csg-based-distributed-firewall