Query Optimization in a Parallel Computer System to Reduce Network Traffic

Abstract

An apparatus and method for a database query optimizer to optimize a query that uses multiple networks. The query optimizer optimizes a query to reduce network traffic on a network or node that is overloaded or above an established parameter in a node/network attribute table. The query optimization to reduce network traffic may result in a sub-optimal query in other respects such as execution time. The result is a query optimizer that rewrites or optimizes a query to execute on multiple nodes or networks to reduce traffic on a network or node according to the loading characteristics and assigned attributes of a node or network.

1 . A computer apparatus comprising:
 a plurality of nodes each having a memory and at least one processor;   a database residing in the memory;   a plurality of networks connecting the plurality of nodes;   a query residing in the memory;   a query optimizer residing in the memory and executed by the at least one processor, wherein the query optimizer creates an optimized query that reduces network traffic over at least one of the plurality of networks.   
     
     
         2 . The computer apparatus of  claim 1  wherein the optimized query that reduces network traffic over at least one network of the plurality of networks is sub-optimal in performance. 
     
     
         3 . The computer apparatus of  claim 1  wherein the optimized query uses a copy node and a different network. 
     
     
         4 . The computer apparatus of  claim 1  further comprising node/network attributes that are setup by a system administrator and used by the query optimizer to determine whether to use multiple networks to optimize the query. 
     
     
         5 . The computer apparatus of  claim 4  wherein the node/network attributes comprise a node/network identification (ID), and an importance. 
     
     
         6 . The computer apparatus of  claim 1  further comprising a service node connected to the plurality of nodes that controls the plurality of nodes, and a network monitor that periodically monitors the plurality of networks to determine network loading. 
     
     
         7 . The computer apparatus of  claim 1  further comprising a query governor that is employed depending on the network loading to determine whether and how to execute the query where the query can not be re-optimized to avoid an overloaded network. 
     
     
         8 . The computer apparatus of  claim 1  further comprising a network file that is used by the query optimizer and wherein the network file contains network file information selected from the following: network ID, a timestamp, current utilization, future utilization, availability, latency and retransmits. 
     
     
         9 . A computer implemented method for optimizing a query on a parallel computer system comprising the steps of:
 receiving a query to the database;   optimizing the query;   determining the query utilizes multiple networks;   determining whether any of the multiple networks are overloaded;   re-optimizing the query to reduce traffic on an overloaded network; and   executing the re-optimized query.   
     
     
         10 . The computer implemented method of  claim 9  wherein the step of re-optimizing the query to reduce network traffic creates a query that is sub-optimal in performance. 
     
     
         11 . The computer implemented method of  claim 9  further wherein the step of determining whether any of the multiple networks are overloaded uses the node/network attributes that are setup by a system administrator by the query optimizer to determine whether to use multiple networks to optimize the query. 
     
     
         12 . The computer implemented method of  claim 11  wherein the node/network attributes comprise a node/network identification (ID), and an importance. 
     
     
         13 . The computer implemented method of  claim 9  further comprising a query governor that is employed depending on the network loading to determine whether and how to execute the query where the query can not be re-optimized to avoid an overloaded network. 
     
     
         14 . A computer apparatus for executing on a parallel computer system with a plurality of compute nodes comprising:
 a query optimizer residing in the memory and executed by the at least one processor, wherein the query optimizer creates an optimized query that reduces network traffic over at least one network of the plurality of networks; and   computer-readable medium in which computer instructions are stored, which instructions, when read by a computer, cause the computer to perform the steps of the query governor.   
     
     
         15 . The computer apparatus of  claim 14  wherein the optimized query that reduces network traffic over at least one network of the plurality of networks is sub-optimal in performance. 
     
     
         16 . The computer apparatus of  claim 14  wherein the optimized query uses a substitute node and a different network. 
     
     
         17 . The computer apparatus of  claim 14  further comprising node/network attributes that are setup by a system administrator and used by the query optimizer to determine whether to use multiple networks to optimize the query. 
     
     
         18 . The computer apparatus of  claim 17  wherein the node/network attributes comprise a node/network identification (ID), and an importance. 
     
     
         19 . The computer apparatus of  claim 14  further comprising a service node connected to the plurality of nodes that controls the plurality of nodes, and a network monitor that periodically monitors the plurality of networks to determine network loading. 
     
     
         20 . The computer apparatus of  claim 14  further comprising a query governor that is employed depending on the network loading to determine how to execute the query where the query can not be re-optimized to avoid an overloaded network.