Please explain the BYNET, Teradata's interconnect component.

The Teradata BYNET interconnect differs from other communication technology on the market (Ethernet, for example) by moving data from one location to another on the network. The network bandwidth is a function only of the equipment purchased.

A gigabyte Ethernet can move 1 GB regardless of how many computers are attached to it: More computers mean more contention and slower performance for all on the network. The BYNET works like a phone switch, quite different from the typical network. Its switched "folded banyan" architecture delivers additional network bandwidth for each node added to the configuration. Each connection to a node delivers 120 MB/sec. A 2-node system has a 240 MB/sec. interconnect; 4 nodes, 480 MB/sec.; 8 nodes, 960 MB/sec.; 100 nodes, 12 GB/sec. The current BYNET implementation can support up to 512 nodes (1 to 4 PB,) but the design is capable of 4,096 nodes (8 to 32 PB) if a customer is ready to solve a problem of that magnitude.

In addition to delivering fully scalable point-to-point bandwidth, the BYNET implements specialized functions not offered by other networking solutions. It can broadcast-deliver a single message-to some or all of the nodes in the MPP configuration. There are many database functions that need to be performed on all nodes at once. With broadcast, the database has to send and manage only one message and one response, lowering the cost and increasing the performance. The BYNET guarantees delivery of every message and ensures that broadcasts get to every target node. So the database isn't plagued by communication errors or network failures and does not have to pay the price of acknowledgements or other error-detection protocols.

The BYNET implements global semaphores so the database can coordinate operations that span some or all nodes. "First done, last done" and counting semaphores implemented at a very low protocol level make these coordination functions efficient and simple. When it is time to return results to a user or application, the BYNET supplies the merge function. Unlike other parallel databases that pull the result set together in one node to deliver the appropriately sorted result, Teradata leaves each portion of the result set on the node where it was created, sorted locally. The merge function collates just enough of the result to fill the user's buffer with result records, then waits for the user to request more rows. Merge makes returning results a scalable operation regardless of their size.
The BYNET performs all of these functions using low-level communication protocols. It is a circuit-switched network, so the large messages that the database sends get through quickly. Higher-level connections are not required; each message simply has a destination address identifying the one or more nodes to which it will be delivered. This messaging architecture eliminates the cost and time required to create, manage and destroy connections in a protocol such as TCP/IP, a cost that escalates exponentially (non-scalable) as nodes are added to the system.