In S-AODV, routing table is maintained at the sink node. This protocol consists of set-up phase and steady phase. In set-up phase, connection is established between the internal nodes and the sink node via optimized path. In steady phase data transfer takes place to and from the destined node through optimized path. Query packets are sent from sink node to the internal nodes. The destined node replies back to the query.
Advantages of S-AODV is minimized power consumption, extended network life time and load balancing in the network. The delay and energy consumption in the network while forwarding packets is minimised to an extent.
In HIERAR, nodes are grouped together to form a multi-level hierarchy of clusters. At level 0, a singleton node itself forms one cluster. At level 1 neighbouring singleton clusters are logically grouped and this procedure continues until some top level clusters are logically grouped together to form top level cluster that covers the whole network. Each cluster has a cluster head assigned. A level-i cluster is advertised to Ri hops away from its head. Level-i cluster is denoted using the notation Cix where ‘x’ is the cluster head.
Each node is identified by a unique identifier. For example, if a node is denoted as P.O.L then it indicates that a node belongs to level 0 cluster with ‘P’ as cluster head, level-1 cluster with node
‘O’ a cluster head and level 2 cluster with cluster head ‘L’. Fig. 4.1 depicts the HiLow routing mechanism.
A node’s routing table contains entries for all the cluster heads from which it receives advertisement. Each entry consists of level and identifier of head, as well as the link-layer address of the next-hop neighbour on the shortest path to the cluster head and the length of this path.
Routing takes place hierarchically. If suppose node ‘C’ labelled as C.A.L wants to send packet to node ‘P’ labelled a P.O.L then node ‘C’ first routes the packet to top-level cluster head of node ‘P’ that is node ‘L’. Node ‘L’, then searches for next decreasing level cluster head in its routing table and routes the packet to node ‘O’ which in turn routes the packet to node ‘P’. So when the packet arrives, the forwarding node will always look for the top-level cluster of the destination node in the routing table entry.
4.4. Improved Hi-Low:
HiLow makes the task of routing too simple by assigning 16-bit short address dynamically in order to establish a hierarchical tree between parent and its children. But this routing protocol suffers from performance degradation as a node tries to communicate with its nearby node over several hops. Improved HiLow boosts the performance of HiLow.
The current node, that receives the packet, identifies its parent through this equation: AP=[(AC-1)/MC]
AP = parent node address AC = current node address
MC = maximum number of children allowed  indicates the floor operation.
The next hop to which the packet should be forwarded is calculated using following 3 cases: If C is a member of SA: The next hop node is AA(DC+1, D).
If C is a member of SD: The next hop node is AA(DC-1, C).
Otherwise, the next hop node is AA(DC-1, C) Where SA = set of ascendants
SD = set of Descendants
D = Destination node address C = current node address
AA(D, k) is the address of the ascendant node of depth D of the node k
DC = Depth of the current node
The steps followed by this protocol are as follows:
1) When a packet arrives at the current node C, it checks whether the packet is destined for itself. If so, then it delivers the packet to its upper layer. If not, it checks whether destination node is a member of SA, if so the next hop is AA(DC+1,D). If not, it checks whether it is a member of SD, if so the next hop is AA(DC-1,D). If not, it continues with the next step.
2) C broadcasts Hello messages in its
Personal Operating Space to get information about Ak and Dk, where Ak is the address of node k and Dk is the depth of node k. It checks whether D is a member of SAk, if so it forwards the packet to D. If not, it checks whether D is a member of descendant of SAk, if so it forwards the packet to one having maximum depth. If not, it checks whether D is a ascendant of SAk, if so it chose the node having minimum depth and forwards the packet to it. Otherwise, it continues with step 3.
3) C calculates all the sibling node of D and checks whether it is member of SAk, if so it forwards the frame to that sibling. If not the next hop is calculated as AA(DC-1,C)
4.5. Dymo Low (Dynamic MANET On Demand for 6LoWPAN Routing):
The DYMOlow protocol provides an effective and simple to implement routing protocol based on AODV.DYMO performs route discovery and maintenance by using RREQ, RREP and RERR messages. It operates on top of IP layer not on the link layer. DYMO protocol cannot be applied directly in 6LoWPAN routing due to its increased memory and power consumption. DYMOLow works in the adaptation layer. It reaches all the nest hop neighbours by broadcasting RREQ messages. It limits the size of RREQ to IEEE 802.15.4 frame size so that it avoids fragmentation. The intended node replies with RREP message. It uses the link quality information to find the best route. It does not use the Hello message instead it uses link layer facility like acknowledgement to get information about its neighbour nodes.
4.6. Extended Hierarchical routing:
During the time of node failure hierarchical routing does not provide path recovery. Extended hierarchical routing provides a path recovery mechanism for 6LoWPANs. This routing protocol adds the Neighbour Replace Parent (NRP) and the Neighbour Added Child (NAC) entry to an existing routing table as a part of path recovery mechanism. NRP does not point to the parent node, but it points to the upstream node to which the packet is delivered. NAC points to the child of the new parent node. For example, suppose node a is down as shown in the figure 4.2.
Now, downstream node 1 cannot transmit packets to the upstream nodes through node a. Here node d is assigned as the NRP of node 4. So node 4, when it does not receive reply from its parent node, assigns a new parent node to node d. After the child node’s NRP is set to new parent node and new parent node’s NAC is set to the child, packet delivery occurs through NAC and NRP.
4.7. Enhanced Location Based Routing Protocol (ELBPR):
This protocol uses both link quality and the distance between nodes as the routing metric. In LOAD, due to the repeated broadcast of Route Request messages (RREQ) during route discovery process, energy consumption increases and also when there is a link failure, route has to be established again as there is only one source. But ELBPR reduces this network overhead as there exists more than one route to the sink. The next best path will be found by the product of LQI (Link Quality) and distance between the nodes (D).
ER (Edge Router) broadcasts its location information in the network. Source node sends RREQ after sensing data. Only nearby LERs (Local Edge Router) that listens to the RREQ replies by unicast RREP message. Using the information sent by LERs, source node constructs neighbour routing table. Depending on maximum distance between LERs and LQI information, the best LER is chosen. Each node maintains a routing table and a neighbour table. Routing table contains location information of ER, address of ER, source address and source location. Neighbour table consists of location of LERs, their address and LQI using which the best LER is chosen.
4.8. MLOAD (Multipath 6LoWPAN Ad-Hoc on demand Distance Vector Routing):
LOAD repeatedly broadcasts RREQ messages during route discovery phase. MLOAD reduces this network overhead by using multiple alternative paths. During route discovery phase it finds multiple paths and when one main route fails it uses the best alternative path for the transmission of data. It broadcasts RREQ messages during route discovery phase and when an intended node receives the RREQ message it replies with RREP message. This protocol also reduces the power consumption comparatively.
5. Comparison of different routing protocols:
Routing protocols described above are compared depending many routing metrics such as energy consumption, memory usage, scalability, mobility, local repair etc.
Energy consumption: all the routing protocols described above consume less energy. MLOAD and ELBPR consume comparatively less energy to LOAD.
Scalability: Hi LoW, Extended Hi Low, Improved Hi Low and ELBPR are highly scalable when compared to other routing protocols.
Hello message: Hello messages are used Hi Low and Extended Hi-Low, S-AODV and Improved Hi-Low for neighbour discovery.
Memory usage: memory usage in LOAD, S-AODV, Improved Hi-Low, MLOAD and DYMO-Low are little high when compared to other routing protocols described.
Mobility: Hi-Low, extended Hi-Low, improved Hi-Low and SAODV are useful in static networks;
where as the rest of the protocols described above are used in dynamic networks.
Routing delay: routing delay is low in all the protocols except DYMO-Low.
Location information: only ELBPR uses the location information for routing.
Type of routing: only S-AODV and LOAD use adaptation layer for routing. So S-AODV and LOAD are mesh under routing. The rest of the routing protocols are route over routing protocols.
A survey has been done on 6LoWPAN routing protocols. This paper describes routing mechanism in 6LoWPAN, routing requirements, its parameters and its types. It also explains various routing protocols such as LOAD, S-AODV, DYMOLow, HiLow, improved HiLow, Extended HiLow , M-AODV and ELBPR. In the end comparison has been made between these protocols. HiLow requires less memory, provides scalability. DYMOLow does not use hello messages thus reducing energy consumption. But it has higher routing delay. S-AODV reduces power consumption and provides network life time extension. Improved HiLow and Extended HiLow provides are more efficient than HiLow. ELBPR overcomes the disadvantages of LOAD and uses location information to find the best route. So, according to the requirements of 6LoWPANs, appropriate routing protocols are chosen that performs well in the established environment.
 Problem Statement and Requirements for 6LoWPAN Routing draft-ietf-6lowpan-routing-requirements-10
 Enhanced location based routing protocol for 6lowpan by m.rehenasulthana, p.t.v bhuvaneswari, n.rama
 6LOWPAN, the wireless embedded Internet byZach Shelby and Carsten Bormann, Wiley publications
 On hierarchical routing in wireless sensor networks by Konrad Iwanicki and Maarten van Steen
 Improved hierarchical routing over 6lowpan by Hong Yu and Jingsha He
 Dynamic MANET On-demand for 6LoWPAN (DYMO-low) Routing draft-6851 ontenegro-6lowpan-dymo-low-routing-03
 J. Chang, T. Chi, H. Yang, and H. Chao, “The 6LoWPAN Ad-hoc on demand distance vector routing with multi-path scheme,” in Proceedings of the IET International Conference on Frontier Computing. Theory, Technologies and Applications, pp. 204–209, August 2010
 Z. Cao and G. Lu, “S-AODV: Sink routing table over AODV routing protocol for 6LoWPAN,” in Proceedings of the 2ndInternational Conference on Networks Security, Wireless Communicationsand Trusted Computing (NSWCTC ’10), vol. 2, pp.340–343, June 2010 C. S. Nam, H. J. Jeong, and D. R. Shin, “Extended hierarchicalrouting over LoWPAN,” in Proceedings of the 4th InternationalConference on Networked Computing and Advanced Information Management (NCM ’08), pp. 403–405, Gyeongju,Korea, September 2008
 Route-over vs Mesh-under Routing in 6LoWPAN by Aminul Haque Chowdhury ,Muhammad Ikram, Hyon-Soo Cha, Hassen Redwan, S.M. Saif Shams, Ki-Hyung Kim, Seung-Wha Yoo
 6LoWPAN overview, “www.mindteck.com”
 Extending IP to Low-Power, Wireless Personal Area Networks by Jonathan W. Hui and David E. Culler
 Implementation and evaluation of Multi-hop routing in 6LoWPAN by Alessandro Ludovici, Anna Calveras
 Connecting Low-Power and Lossy Networks to the Internet by JeongGil Ko and Andreas Terzis, Stephen Dawson-Haggerty, David E. Culler, Jonathan W. Hui and Philip Levis