Results 1 -
6 of
6
Time-Shift Scheduling – Fair Scheduling of Flows in High-Speed Networks
- IEEE/ACM Transactions on Networking
, 1998
"... We present a scheduling protocol, called Time-Shift scheduling, to forward data packets from multiple input flows to a single output channel. Each input flow is guaranteed a predetermined forwarding rate and an upper bound on packet delay. The protocol is an improvement over existing protocols becau ..."
Abstract
-
Cited by 19 (3 self)
- Add to MetaCart
We present a scheduling protocol, called Time-Shift scheduling, to forward data packets from multiple input flows to a single output channel. Each input flow is guaranteed a predetermined forwarding rate and an upper bound on packet delay. The protocol is an improvement over existing protocols because it satisfies the properties of low delay, fairness, and efficiency, while existing protocols fail to satisfy at least one of these properties. In Time-Shift scheduling, each flow is assigned an increasing timestamp, and the packet chosen for transmission is taken from the flow with the least timestamp. The protocol features the novel technique of time shifting, in which the scheduler's real-time clock is adjusted to prevent flow timestamps from increasing faster than the real-time clock. This bounds the difference between any pair of flow timestamps, thus ensuring the fair scheduling of flows. 1.
Balanced routing
- IEEE Proceedings of the International Conference on Network Protocols
, 1997
"... The distance vector routing protocol satisfies the following property. If this protocol is used to route a sequence of data messages from a source to a destination, then these data messages will follow the same shortest-distance path from the source to the destination. In this paper, we show how to ..."
Abstract
-
Cited by 8 (2 self)
- Add to MetaCart
The distance vector routing protocol satisfies the following property. If this protocol is used to route a sequence of data messages from a source to a destination, then these data messages will follow the same shortest-distance path from the source to the destination. In this paper, we show how to modify this protocol, without adding new messages, in order to satisfy the following load balancing property. If the modified protocol is used to route a sequence of data messages from a source to a destination, then these messages will be uniformly distributed over all k-monotonic paths from the source to the destination, i.e., over all paths whose distance to the destination never increases at each hop, and may remain constant in at most k hops. In particular, by choosing k to be zero, the modified protocol distributes the data messages over all shortest-distance paths from the source to the destination. 1.
Propagated Timestamps: A Scheme for the Stabilization of Maxmum-Flow Routing Protocols
- In WSS97 Third Workshop on Self-Stabilizing Systems
, 1997
"... We present a distributed protocol for maintaining a maximum flow spanning tree in a network, with a designated node as the root of the tree. This maximum flow spanning tree can be used to route the allocation of new virtual circuits whose destination is the designated node. As virtual circuits are a ..."
Abstract
-
Cited by 6 (3 self)
- Add to MetaCart
We present a distributed protocol for maintaining a maximum flow spanning tree in a network, with a designated node as the root of the tree. This maximum flow spanning tree can be used to route the allocation of new virtual circuits whose destination is the designated node. As virtual circuits are allocated and removed, the available capacity of the channels in the network changes, causing the chosen spanning tree to lose its maximum flow property. Thus, the protocol periodically changes the structure of the spanning tree to preserve its maximum flow property. The protocol is self-stabilizing, and hence it tolerates transient faults. Furthermore, it has the nice property that, while the structure of the tree is being updated, no loops are introduced, and all nodes remain connected. That is, the tree always remains a spanning tree whose root is the designated node. 1.
The Request-Reply Family of Group Routing Protocols
- ACM Transactions on Computers
, 1998
"... We present a family of group routing protocols for a network of processes. The task of these protocols is to route data messages to each member of a process group. To this end, a tree of processes is constructed in the network, ensuring each group member is included in the tree. No processing or sto ..."
Abstract
-
Cited by 4 (4 self)
- Add to MetaCart
We present a family of group routing protocols for a network of processes. The task of these protocols is to route data messages to each member of a process group. To this end, a tree of processes is constructed in the network, ensuring each group member is included in the tree. No processing or storage overhead is required for processes not included in the tree. The overhead of processes in the tree consists solely of the periodic exchange of request/reply messages with their parent. To choose the processes that constitute the tree, we take advantage of the existing unicast routing protocol in the network. In addition, our family of group routing protocols distinguishes itself from other group routing protocols in three ways. First, the protocols are proven correct. Second, the protocols preserve the integrity of the group tree as it adapts to changes in the unicast routing tables, even in the presence of temporary unicast routing loops. Third, data messages are propagated along the entire group tree, even while the tree adapts to changes in the unicast routing tables. 1.
Universal Timestamp Scheduling for Real-Time Networks
- Computer Networks
, 1999
"... Consider a network of computers interconnected by point-to-point communication channels. Before generating network packets, each source in the network reserves a fraction of the packet rate of each output channel in the path to its destination. We define a family of scheduling protocols, called Univ ..."
Abstract
-
Cited by 3 (1 self)
- Add to MetaCart
Consider a network of computers interconnected by point-to-point communication channels. Before generating network packets, each source in the network reserves a fraction of the packet rate of each output channel in the path to its destination. We define a family of scheduling protocols, called Universal Timestamp-Scheduling, to forward packets in this network such that all members of the protocol family provide the same upper bound on packet delay as Virtual Clock scheduling. That is, the packets from a source will exit the output channel of a computer no later than the time they would exit an output channel whose rate equals the source's reserved rate and whose input is exclusively the packets from this source. The protocol family is called universal because it encompasses a wide variety of protocols. To show this, we prove that many scheduling protocols in the literature are members of the protocol family, and thus provide the above guarantee. In addition, we show that the protocols in the literature have only considered one side of the spectrum of possible scheduling protocols, and that there is another side of the spectrum that deserves attention and remains to be investigated. 1.
Group Routing without Group Routing Tables: An exercise in Protocol Design
- University of Houston
, 1996
"... We present a group routing protocol for a network of processes. The task of the protocol is to route data messages to each member of a process group. To this end, a tree of processes is constructed in the network, ensuring each group member is included in the tree. To build this tree, the group rout ..."
Abstract
-
Cited by 2 (2 self)
- Add to MetaCart
We present a group routing protocol for a network of processes. The task of the protocol is to route data messages to each member of a process group. To this end, a tree of processes is constructed in the network, ensuring each group member is included in the tree. To build this tree, the group routing protocol relies upon the local unicast routing tables of each process. Thus, group routing is accomplished by composing two protocols: an underlying unicast routing protocol, whose detailed behavior is unknown but its basic properties are given, and a protocol that builds a group tree based upon the unicast routing tables. The group routing protocol is developed in three steps. First, a simple protocol is obtained, and is proven correct. Then, the protocol is refined twice. Each refined protocol improves upon its predecessor by satisfying all of the predecessor's properties plus some additional stronger properties. The final protocol has the property of adapting the group tree to changes in the unicast routing tables without compromising the integrity of the group tree, even in the presence of unicast routing loops. 1.

