Wormhole Switch Download
Ann Iacobucci
With the JUC400 there is no software to install or lengthy settings to adjust. Simply plug the wormhole cable into device A and device B and you're ready to go! Drag and drop or copy/paste files easily and quickly across Windows to Windows or Mac to Mac and even direct transfers between Mac and Windows.
Switching is a more appropriate term than routing, as "routing" defines the route or path taken to reach the destination.[2][3] The wormhole technique does not dictate the route to the destination but decides when the packet moves forward from a router.
In wormhole switching, each buffer is either idle, or allocated to one packet. A header flit can be forwarded to a buffer if this buffer is idle. This allocates the buffer to the packet. A body or trailer flit can be forwarded to a buffer if this buffer is allocated to its packet and is not full. The last flit frees the buffer. If the header flit is blocked in the network, the buffer fills up, and once full, no more flits can be sent: this effect is called "back-pressure" and can be propagated back to the source.
The name "wormhole" plays on the way packets are sent over the links: the address is so short that it can be translated before the message itself arrives. This allows the router to quickly set up the routing of the actual message and then "bow out" of the rest of the conversation. Since a packet is transmitted flit by flit, it may occupy several flit buffers along its path, creating a worm-like image.
This behaviour is quite similar to cut-through switching,[5] commonly called "virtual cut-through," the major difference being that cut-through flow control allocates buffers and channel bandwidth on a packet level, while wormhole flow control does this on the flit level.
Wormhole techniques are primarily used in multiprocessor systems, notably hypercubes. In a hypercube computer each CPU is attached to several neighbours in a fixed pattern, which reduces the number of hops from one CPU to another. Each CPU is given a number (typically only 8-bit to 16-bit), which is its network address, and packets to CPUs are sent with this number in the header. When the packet arrives at an intermediate router for forwarding, the router examines the header (very quickly), sets up a circuit to the next router, and then bows out of the conversation. This reduces latency (delay) noticeably compared to store-and-forward switching that waits for the whole packet before forwarding. More recently, wormhole flow control has found its way to applications in Network On Chip systems (NOCs), of which multi-core processors are one flavor. Here, many processor cores, or on a lower level, even functional units can be connected in a network on a single IC package. As wire delays and many other non-scalable constraints on linked processing elements become the dominating factor for design, engineers are looking to simplify organized interconnection networks, in which flow control methods play an important role.
An extension of worm-hole flow control is Virtual-Channel flow control, where several virtual channels may be multiplexed across one physical channel. Each unidirectional virtual channel is realized by an independently managed pair of (flit) buffers. Different packets can then share the physical channel on a flit-by-flit basis. Virtual channels were originally introduced to avoid the deadlock problem, but they can be also used to reduce wormhole blocking, improving network latency and throughput. Wormhole blocking occurs when a packet acquires a channel, thus preventing other packets from using the channel and forcing them to stall. Suppose a packet P0 has acquired the channel between two routers. In absence of virtual channels, a packet P1 arriving later would be blocked until the transmission of P0 has been completed. If virtual channels are implemented, the following improvements are possible:
A mix of source routing and logical routing may be used in the same wormhole-switched packet.The value of the first byte of a Myrinet or SpaceWire packet is the address of the packet.Each SpaceWire switch uses the address to decide how to route the packet.[7]
If the first byte of an incoming SpaceWire packet is in the range 1 to 31,it indicates the corresponding port 1 to 31 of the Spacewire switch.The SpaceWire switch then discards that routing character and sends the rest of the packet out that port.This exposes the next byte of the original packet to the next SpaceWire switch.The packet sender may choose to use source routing to explicitly specify the complete path through the network to the final destination in this fashion.[7]
If the address (the first byte) of an incoming SpaceWire packet is in the range 32 to 255,the SpaceWire switch uses that value as an index into an internal routing table that indicates which port(s) to send the packet and whether to delete or retain that first byte.[7]
In data communications, wormhole switching a flow control technique where large data frames or packets are partitioned and then transmitted. When a switching device (a bridge or a switch) receives a data packet, it partitions thepacket into small parts called flow control units or flits. The flits are transmitted one by one instead of the whole packet. Also called wormhole flow control, wormhole switching is subtype of flit-buffer flow control methods and is basedupon fixed links.
After joining UTSA in 1991, I initiated work on the design of wormhole-switchedmulticomputers networks and routing algorithms. The main results are design of adaptiverouting algorithms compatible the with buffered wormhole switching used in IBM's SP-2network design, fault-tolerant routing techniques for mesh networks to handle faultson-the-fly with local fault information, and multicast routing methods.
The new adaptive algorithms proposed work for any network topology. Itis shown that these algorithms give better throughputs and can be implemented with lessbuffer space than previously proposed adaptive wormhole routing algorithms. Significantpublications based on this work include papers in IEEE TPDS '96 and ISCA '93.
The multicast routing work analyzed a new form of deadlocks that occurwith multicast communication in wormhole networks and proposed techniques to remedy thesame. Also, the commonly used e-cube routing is adapted for multicast routing and is shownusing extensive simulation studies that its performance is superior to many known andcontemporary multicast techniques. This work was published in papers in IEEE TPDS'98 and SPDP '94.
Because of increasing electronic commerce and voice transmissions on the Internet, fast switches that can sustain multi-gigabit and terabit throughputs will be needed. These switches will need to support fast packet switching, stringent quality of service (QoS) guarantees, multicasts, and common standards. Furthermore, fault-tolerance capabilities will become even more crucial as more and more business is transacted on the Internet. I am currently investigating high-speed switch designs that can be scaled from gigabits/sec to terabits/sec throughputs and support a variety of traffic patterns (best-effort, QoS, multicasts). I am applying some of my earlier fault-tolerant routing techniques to handle faults on-the-fly with local information, while avoiding major network outages. A couple of results in this direction were presented at ICC '99 and HiPC '03.
df19127ead