Scalable Content-centric Routing for Hybrid ICN
0
Citation
15
Reference
10
Related Paper
Abstract:
Hybrid Information-Centric Networking (hICN) is an incrementally-deployable information-centric networking architecture that is built on top of IPv6. In hICN, application-level identifiers are directly used to route interest packets (i.e., request for content) to fetch a copy of the desired content/data from any location. However, following the Internet Protocol conventions that require storing pre-computed routing/forwarding state for all prefixes in the routers raises scalability concerns, especially at the inter-domain level. Here we consider instead the other extreme; i.e. on-demand routing computation for content name prefixes when interest packets arrive at the router. Following this approach, we propose a centralized routing service within a domain that keeps a mapping between hICN name prefixes and locators (i.e., routable addresses) to hICN routers. Once a locator is received, an hICN router forwards an interest packet towards the intended destination using segment routing. We evaluated the proposed solution through a real testbed implementation in order to demonstrate that the performance is equivalent to typical hICN forwarding, while offering a scalability solution.Keywords:
Policy-based routing
Testbed
Loose Source Routing
Source routing
So that the routers forward an IP packet with his destination, they are running a forwarding decision on an incoming packet to determine the packet’s next-hop router. This is achieved by looking up the longest matching prefix with a packet destination address in the routing tables. Therefore, one major factor in the overall performance of a router is the speed of the IP address lookup operation due to the increase in routing table size, in this paper, a new IP address lookup algorithm based on cache routing-table is proposed, that contains recently used IP addresses and their forwarding information to speed up the IP address lookups operation in the routers. We evaluated the performance of our proposed algorithm in terms of consultation time for several sets of IP addresses, the results of performance evaluation show that our algorithm is efficient in terms of the lookup speed since search can be immediately finished when the input IP address is found in the cache routing table.
IP forwarding
Loose Source Routing
Source routing
Packet forwarding
Equal-cost multi-path routing
Cite
Citations (4)
IP forwarding
Packet forwarding
Loose Source Routing
Source routing
Triangular routing
Equal-cost multi-path routing
Active networking
Cite
Citations (0)
The paper proposes a novel algorithm for a Network-onChip, which is based on packet switching. Unlike datagram protocol used in large communication network, which can be one of the algorithms in NoC; this paper concentrates on static method of routing of packets. The algorithm described here, uses three different types of packets to achieve communication between various intellectual properties connected to the chip. A packet named route establisher commences the start of packet transfer establishing a fixed route from the source node to the destination node. The packet following route establisher is the data packet, which hops through the same nodes as fixed by the route establisher. When all the data is sent or received, source IP has an option of destroying the link with the destination IP, using route destroyer packet, or might keep it for future communication. A simple prototype using sixteen nodes is used in the design to prove the working of NoC, which further can be expanded to any number of nodes as per the requirement in the design. The proposed algorithm is applicable for a network which is a square mesh topology. The number of nodes in a row should be exactly equal to the number of nodes in a column. The paper also enhances a unique internal architecture of a node in NoC, focusing on the uses of Content Addressable Memory (CAM) as a routing table in a node. General Terms Network-on-Chip, Algorithms, Content Addressable Memory.
Source routing
DSRFLOW
Packet forwarding
Packet generator
Loose Source Routing
End-to-end delay
Cite
Citations (0)
When routing-switch forwarding IP packet, it will check its router table entries for best path to the destination. It's a time-consuming process to realize this function at network layer. With network scale grows, the router table expands much bigger and searching path in routing table becomes a bottleneck. Address cache can be used to speed up the routing process. This is done by routing first packet normally and forwarding subsequent packets directly using the route information found in first packet routing process. To realize this concept, an address cache is needed to cache the route information so that subsequent packets can be flown through the routing-switch directly. ;;;
Equal-cost multi-path routing
Source routing
Packet forwarding
IP forwarding
Loose Source Routing
Policy-based routing
Cite
Citations (0)
Routing domain
Policy-based routing
Source routing
IP forwarding
Equal-cost multi-path routing
Packet forwarding
Cite
Citations (1)
Link state IP protocols like OSPF and IS-IS work upon network state advertisements. Link state packets travel all over the network, advertising recent failure or changes in the network. It is IP router responsibility to execute the complex updating process which includes recalculating shortest paths and updating central and local forwarding tables. Since forwarding tables must be updated for forwarding new arrival packets, forwarding received packets during the updating process is one of the most critical issues. There have been several researches in the way of updating forwarding tables for achieving less packet loss ratio and faster packet forwarding statistics during updating process. This paper presents an updating method based on clustering the traffic which has been recently received by the router. Each cluster represents a set of network prefixes with almost the same priority for being updated. Our experimental results shows how clustering could decrease the packet loss ratio during updating process.
IP forwarding
Packet forwarding
Forwarding plane
Loose Source Routing
Source routing
Link state packet
Cite
Citations (0)
With packet radio networks, the distance between source and destinat.ion nodes typically necessitat$es one or more nodes to relay data to the final destination. Thus, some form of routing must. take place. This paper explains several current network routing algorithms and shows their relevance to packet, radio networks. In addition, current research at. AFIT concerning the development of an automatic rout.ing algorithnl for Air Force Logistics Command’s (AFLC) HF packet radio network is explained. 1 I N T R O D U C T I O N Most, routing algorithms store node address information in tables, which show the next node to rec,eive a packet. These routing algorithms may be stat*ic or dynamic. If tOhe algorithms are static (nonadaptive) the table entries do not change during normal operateion of the network. Dynamic (adaptive) routing algorithms periodically update the tables too reflect. changes in the network’s topology or utilization or both [6]. Since packet1 radio networks typically have changeable connectivitly, the routing mechanism must be capable of updating routing tables. This paper examines three basic t,ypes of adaptive routing algorithms: centralized, distributed, and isolated 171. The final section describes current0 research of an automatic routing algorithm for AFLC’s HF packet. radio network. 2 CENTRALIZED ROUTING A centralized routing algoritlhm requires a routing control center (RCC) to make routing decisions based on information gained from each node within the network. Each node monitors connectivity and delay metrics among neighboring nodes and periodically sends this information to the RCC. The RCC calculates the best1 route (normally in terms of least, delay) and sends each node new routing table information depending* on the nlost. recently measured state of the network. s 105 In most cases, a source node needing to send data packets can notify the RCC of the source and destination. The RCC will respond with a special call request packet called a needle packet, that contains the route which is the most efficient circuit. The route is specified as an ordered set+ of nodes. The source node then sends the needle packet through the network to establish the circuit, and then data packets can follow. Although centralized routing offers a solution to adaptive routing, this technique has disadvantages worth discussing. For one, centralized routing requires a large amount of overhead due to routing information sent between nodes and the RCC. As a result, centralized routing may not, be suitable for some networks operating with limited bandwidth, such as HF packet. radio. Also, large networks having many nodes will require the RCC to perform lengthy calculations to determine optimum routes. Hence, the “optimum” table ent.ries may not be valid if the network topology changes rapidly. 3 DISTRIBUTED ROUTING With distributed routing, each node distributes routing metrics (connectivit.y informatioll, node delays) throughout the network, enabling other nodes to update routing tables. Distributed routing has proven to be very robust wit.h ARPANET (Advanced Research Project Agency Network). Within AR.PANET, each node periodically measures the delay to each node Ahin one transmission hop and puts this information into a status packet. Nodes wiUlin one tlransnrission hop are known as neighbors. The node then transmits the stat.us packet9 to each neighbor, which records the statlus information. Each neighbor repeats t,he delay measuring process, formulates a status packet9 containing local delay informadion as well’as delay information from incoming status packets and sends this status packet. to each of its neighbors. By having all nodes follow this process, each node will eventuall\v have an overall “picture” of the network in terms of node-tonode delays. Each of thb nodes can then determine the route of least delay by referring to the status information received from other nodes. Distributed routing requires a significant amount of overhead with packet radio net,works having highly mobile nodes. Each node must send status information often enough to account for rapid changes in topology due to node movement. Hence, the network must have a considerable amount of bandwidth available or suffer from rather low throughput. Jubin and Tornow explain that. the DARPA (D f e ense Advanced Research Project. Agency) packet radio network (not ARPANET) applies distributed routing techniques by having each node maintain a tier table [2]. The tier table specifies nodes that are one hop away (tier I), two hops away (tier 2), three hops away (tier 3), and so on. The tier table is arranged in a matrix format. The tiers represent. the rows, whereas the tier 1 ent$ries head off the columns. For example, node X could have a tier table as shown in Figure 1. Here, nodes A, B, and C are neighbors of node X; nodes D and E are neighbors of node A; node N and J are neighbors of node D; node F is a neighbor of node B, and
Policy-based routing
Geographic routing
Multipath routing
Source routing
Equal-cost multi-path routing
Hierarchical routing
DSRFLOW
Cite
Citations (0)
As plans for network interconnection develop, the problems of internet routing and addressing become increasingly important. In one popular model of internet addressing, a hierarchical form of network and local (within network) address is used, with the source providing only the destination address while the intermediate network(s) and/or Gateways between networks take care of routing packets to that destination by various paths. This and related techniques requiring some form of routing table and knowledge at intermediate nodes are more fully discussed in [1,2,3]. This paper considers another technique for internet routing in which the source of internet packets specifies the complete internet route. When the entire route accompanies each internet packet, no routing decisions or tables are required at Gateways, but the packet format is complicated and overhead increases. In particular, the packet must carry a varying number of intermediate addresses depending on the path and destination [4]. This overhead may be reduced by setting up a fixed route with connection tables [1] when a connection is established.
Source routing
IP forwarding
DSRFLOW
Hierarchical routing
Policy-based routing
Equal-cost multi-path routing
Geographic routing
Loose Source Routing
Cite
Citations (104)
Today's modern tactical operations are going towards the concept of Network Centric Warfare. In such military operations, tactical wireless networks require high demands for robustness, responsiveness, reliability, availability and security. One of the key requirements of the tactical user expectation is to efficiently utilize the available bandwidth. IP packet routing in the tactical wireless networks which are connected via IP radios consume more system bandwidth in order to maintain the up-to-date information. Complex routing protocols have to be implemented in the radio as embedded software modules in order to exchange the IP routing information. This will consume the system bandwidth thereby throughput will be reduced. Moreover, hop by hop routing method will reconstruct the entire IP packet by changing the destination MAC address in every hop thereby packet processing time will be increased. In this paper, we have proposed two simple and efficient IP packet forwarding schemes for multi-hop tactical wireless networks which are connected via radios. Both schemes make use additional parameters such as unique radio identifier and associated IP Address/MAC address table during the topology discovery process using the link state routing algorithm. After successful topology discovery, each wireless node in the network will get to know the entire network topology in terms of a table containing the radio connectivity list and each radios associated LAN IP address/MAC address. This network topology table will be used in during the packet forward decision making process. The associated IP/MAC address to node Id mapping table will be used to find the destination radio Id from the incoming IP packet. The forwarding radio nodes will examine only the radio header portion of the received IP packet instead of analyzing & reconstructing the entire IP packet thereby packet processing time will be reduced at each intermediate wireless node. There by packet processing time will be reduced. The proposed approach also supports for forwarding of broadcast and multicast IP packets to enable group communication among the IP networks which are connected via the IP radios. Overhead for both of the proposed approaches are very minimal and consume less bandwidth.
IP forwarding
Packet forwarding
Source routing
Loose Source Routing
Cite
Citations (2)