Tuesday 6 August 2013

Network Management System in Mobile Ad Hoc network


Abstract— a mobile Ad Hoc wireless network (MANET) is a network established for special, often extemporaneous service customized to applications. MANETs are self-organizing network characterized by low bandwidth, variable link state and by constant topology change subject to random changes in formation and deformation, posing many challenges in Network management.
This paper describes an overview of Network Management system core features for large mobile Ad Hoc network. 

1. Introduction


Mobile Ad hoc Network (MANET) is a self con-figuring multi-hop wireless network of mobile nodes without any fixed or pre-existing infrastructure. Nodes within the wireless range of each other can communicate directly. Nodes outside each other’s wireless range must communicate indirectly, using a multi-hop route through other nodes in the network. Nodes in MANET are free to move randomly and organize themselves arbitrarily. The topology may change rapidly and unpredictably. Such a network may operate in a standalone fashion, or may be connected to the larger Internet. MANET has numerous applications such as sensor networks, disaster relief, military operations, business and home applications.

Utilizing clustering algorithm to construct hierarchical topology in MANET may be a good method to solve some of the issues currently faced in the MANET network. Clustering of nodes in MANET could reduce overhead, flooding, collision and duplicate IP address in communication and make the network topology more stable. In clustering algorithm the whole MANET will be divided into several clusters so that, a node joins or quits the network will only affect one cluster not the whole network. In hierarchical topology, each cluster will have Cluster Head, Cluster Gateway and Cluster member node. The cluster head is usually one hop away from the other cluster member. In this paper we assume that each cluster has dis-joint set of IP set pool, so that duplicate IP address is completely avoided across the MANET network

Network management plays a vital role in the well-being of any communications network. It becomes extremely important and crucial, in order to keep the whole network and application work properly.

MANETs present several constraints; they are characterized by a dynamic environment and the scarcity of resources. These features make their management rather difficult and a challenging task. Furthermore, the traditional network management approaches become impractical for these networks.

To simplify the task of management, ANMP proposes two clustering algorithms: graph-based algorithm and geographical based algorithm. We are considering how to build the Network Management system application using only the graphical based algorithm.





Cluster: A cluster is a logical entity consisting of a group of mobile nodes. Usually there will be a limitation on the number of nodes present in a cluster.

Cluster Head: The cluster head is the backbone of small network called cluster. It maintains two IP tables, one for its cluster nodes and other for neighboring cluster heads. The Cluster node IP allocation table is used to provide unique IP address to its member’s nodes and keep track of the assigned, allocation in progress, reserved and free set of IP addresses. Neighbor Cluster Head IP table will have summary of IP address details of the neighbor cluster heads. Any node in the cluster has the capability to be elected as cluster head, if it satisfies the set of cluster head selection criteria and once elected it will broadcast to entire cluster saying he is the current cluster head. One backup cluster head will be selected during cluster head selection procedure and it will be in standby position and ready to take cluster head job in case of any failure or issues in the current cluster head. This backup cluster head will behave as normal cluster member node, before taking up cluster head job.  There will be a periodic synchronization happens between the cluster head and backup cluster head. The cluster head in a cluster can communicate to other neighbor cluster heads in the MANET either directly or via Cluster gateway

Cluster Gateway: The cluster gateway is used for inter-cluster communication. Suppose if a node present in the cluster wants to communicate to other node present in another or neighbor cluster it has to go via cluster gateway. Any node in the cluster has the ability to become cluster gateway if it satisfies the cluster gateway selection criteria (Simple criteria for cluster gateway selection could be receiving signal from two or more cluster head or nodes). Cluster gateway will be selected by the cluster head based on the selection criteria. Cluster gateway plays an important role in stability of MANET network. There is a possibility for the cluster head to behave as cluster gateway also, if it satisfies cluster gateway criteria also.

Cluster Node: After joining the cluster (i.e., Network Entry, Authentication and IP Assignment), a non cluster node can become a cluster node and become an active member of that cluster. The member cluster node will be able to communicate with other members of the cluster either by one hop or by multiple hops. If the member node wants to communicate with node present in other cluster it has to be via cluster gateway. All the member nodes actively participate in cluster head and gateway selection. The cluster head and cluster gateway is a part of cluster node.

Non cluster node: A node which is standalone and not connected to any other cluster is called non-cluster node.
Different clustering algorithms can be used for the dynamic creation of clusters and the election of cluster heads. At lowest level (level 1), we have many clusters with each clusters having dis-joint pool of IP set. Level 2 and level 3 clusters are generated recursively; new cluster heads are elected at each level and become members of the higher level cluster.

In order to have a management protocol that is message efficient, a hierarchical model for data collection is appropriate, since intermediate levels of the hierarchy can collate data (possibly producing a digest) before forwarding it to upper layers of the hierarchy. A problem, however, with utilizing a hierarchical approach in ad hoc networks is the cost of maintaining a hierarchy in the face of node mobility. A good tradeoff we observed is to use a multi-level hierarchical architecture. The lowest level of this architecture consists of individual managed nodes called agents. Several agents (that are close to one another) are grouped into clusters and are managed by a cluster head. The structure of the clusters is dynamic. Thus,  as nodes move about, the number and composition of the clusters changes. Similarly, the nodes serving as cluster heads also changes over time


2. Difficulties in Managing MANET:         
·         The topology of an ad hoc network may be quite dynamic. So a network management system must adapt to changes in topology and be fairly robust to intermittent link and node failures, changes in link quality and network partitioning. Thus, the management station needs to collect connectivity information from nodes periodically.
·         The management plane must have the ability to quickly react to all changes in node state and report the change to a higher layer management application that may be located at a remote node within a certain time. Furthermore, most nodes in ad hoc networks run on batteries. Thus, we need to ensure the management overhead is kept to a minimum so that energy is conserved. As a result, we need to balance the management protocol performance against its overhead.
·         Ad hoc network nodes can range in complexity from simple sensors to fully functional computers such as laptops. Thus, the management protocol needs to function in very heterogeneous environments. At the same time, the management protocol should be compatible with the variety of hosts and the current management protocols because there is an obvious need for interoperability
·         Ad hoc networks are employed for a diverse set of applications. These applications have different security requirements, and the management protocol should be able to adapt easily to the characteristics of the situation.
·         Scalability: The network management system must scale to be able to manage a mobile ad hoc network consisting of thousands of nodes.



3. Proposed Network Management Solution:

Network Management System applications in MANETs include network monitoring, configuration and control of network and node resources (like battery life, bandwidth, etc.). It is assumed that network entry of nodes into the clusters will be always in a secured manner. In this architecture the NMS station will be connected only to the Cluster Head and not to all the nodes. Cluster head will persists all the nodes in that particular cluster and communicates the same to the NMS station.  Our proposed solution of NMS will manage the MANET using ANMP protocol.




3.1. ANMP Architecture:

Ad hoc network management protocol (ANMP) is fully compatible with simple management protocol, version 3 (SNMPv3) and uses the same protocol data units (PDU’s) for data collection. The protocol also implements sophisticated security mechanisms that can be fine-tuned to meet specific requirements.



There are three core functional areas of a MANET NMS.


Fault Management: Detect, isolate, and notify faults encountered in the Mobile Ad Hoc network.
Configuration Management: Configuration aspects of network devices such as parameter configurations.
Performance Management: Monitor and measure various aspects of performance including QoS so that optimization can be done on the network.

Organizational Model: Centralized mode – We prefer to develop the Network Management System in this mode as this model has got an advantage to simplify the management task by considering a single point of control.
 


Topology: NMS topology feature will request for the operator to discover the nodes based on the IPAddress range for the first time. NMS will discover nodes based on the SNMP credentials provided and all the nodes will be added in NMS under different clustered topological view. NMS topology also gives a high level tactical overview of the Mobile Ad Hoc network by displaying few critical characteristics like Node availability (includes Reachability). Topology feature in NMS needs to display the frequent changes in the network topology. It is assumed that the Cluster head periodically updates the cluster and all its node information through ANMP protocol.


Fault Management:
This is required for the smooth operation of the MANET. Whenever a fault occurs in the network For example: Node dies, Node powered off, the same will be sent as ANMP Traps to the NMS station. It will be displayed to the operator to take necessary actions. Apart from Faults in the network, events also will be triggered by this management from the Cluster Head to the NMS station such as follows:
• a new neighbor in the cluster
• that a previous neighbor and cluster mate has moved
• that it was previously directly connected to the cluster head but is no longer directly connected, or it was
previously not directly connected but is now directly connected;
• a new neighbor, who wants to join the cluster.
All these events will be processed by the NMS station and displayed to the operator immediately.

Performance Management:
This feature is responsible for collecting the network statistics such as memory utilized,  interface availability, network traffic, CPU utilization, network availability, disk space, error rate, packet loss, etc. All these values will be present in the ANMP agent maintained in each node in a cluster, the cluster head collects it and persists. Based on granularity period interested by the operator, NMS station can query the Cluster Head and presents it to the operator as always the users expects same amount of quality of service provided by the wired line networks.

Configuration Management:
This feature is the most critical one in NMS application. The purpose of configuration management is to maintain up-to-date information about the operation and configuration status of nodes in MANET. Specific configuration functions include the management of the radio network, software and hardware management of the network elements, time synchronization, and security operations, where as it depends on what configuration parameters are exposed in the ANMP implemented O&M agents in MANET. ANMP PDU will be built by the NMS station with respective SNMP GET/SET operations and it allows the operator to configure based on the need.  Security level of message (PDU) transfer between the NMS station and the Cluster head is always configurable at both the ends.
The following are the critical parameters NMS operator is interested on: IPAddress details, cluster details, type of node, services offered to the node, Node position in the cluster, velocity, neighbor list and direction. Few parameters will be computed by the node and which are read-only. Either NMS can query the Node’s MIB and retrieve or the node itself can send this information as SNMP Trap PDU to the NMS station. It purely depends up on the node’s agent architecture.










4. Scenarios to be handled in NMS for managing MANET
NMS station is supposed to adhere to the following actions:
  1. NMS has to frequently update the node’s status information such as Node’s graceful exit, abrupt exit & Flapping. Since these operations are quite frequent in MANET, NMS should adhere to all these and also considering NMS should not use much bandwidth & frequent SNMP GET operations as there is serious limitation in the energy & bandwidth. In order to achieve better management result in NMS, Cluster head should update this information on a periodic basis to the NMS station.
  2. Alarm/Event Synchronization: Assuming ANMP is a similar protocol like SNMP, which is implemented over UDP, there is a high chance of message getting lost in the network (because of non-acknowledgement nature, dynamic changes in the Topological connection etc..). In order to overcome this, NMS station should have a feature of synchronizing alarm/event information available with the cluster head/node. This action has to be performed manually by the operator on need basis such as when there is no frequent communication between the nodes in the network.
  3. Heartbeat check: NMS station will perform a reachability check frequently on cluster heads. If cluster head doesn’t respond, NMS will mark it as DOWN in a status table maintained locally by NMS. Heartbeat check’s frequency will be increased for those non-reachable cluster heads and it will update the Topological information after the new cluster head arrives.
  4. Partitioning & merging of clusters: In MANET there can be situation in which two or more disjoint cluster can be joined as a single cluster or vice versa. In order to manage this, a NMS operator can manually change the topological cluster information maintained in the NMS (or) the newly formed cluster head will notify this change to the NMS which NMS has to update in its logical view.
  5. Self-configuration: It is also possible in MANET that a node can automatically configure few configuration parameters without the influence of the operator triggered NMS configuration. In this case, NMS has to adhere to these frequent changes by retrieving the latest/current MIB configuration values based on need.
  6. Centralized NMS server crash: In this paper Centralized mode of NMS station is only considered, if the NMS server crashes, there is no possibility of managing a MANET. When the NMS server comes up, NMS should be capable of updating the network information by configuration synchronization.

5. Conclusion:
In conclusion, we can say that none of the current Network management system available in the market today provides a full fledged MANET network management. In this paper, we have described core functionalities of a NMS station, difficulties in managing a MANET and adhering to MANET’s special scenarios. It is possible to effectively manage a MANET by considering the critical aspects mentioned in this paper.  We have also considered few important scenarios that can occur frequently in MANET which has been take care in this paper. 

Authors: Ramkumar, Mahendar & Kannan
References: Wikipedia, IEEE papers on MANET

Note: Also, i found this is useful 
ANMP: ad hoc network management protocol

http://ieeexplore.ieee.org/xpl/articleDetails.jsp?arnumber=780355

No comments:

Post a Comment