NETWORK CAPACITY PLANNING FOR MULTIPLE INSTANCES OF AN APPLICATION
Data representing application deployment attributes, network topology, and network performance attributes based on a reduced set of element attributes is utilized to simulate application deployment. The data may be received from a user directly, a program that models a network topology or application behavior, and a wizard that implies the data based on an interview process. The simulation may be based on application deployment attributes including application traffic pattern, application message sizes, network topology, and network performance attributes. The element attributes may be determined from a lookup table of element operating characteristics that may contain element maximum and minimum boundary operating values utilized to interpolate other operating conditions. Application response time may be derived using an iterative analysis based on multiple instances of one or more applications wherein a predetermined number of iterations are used or until a substantially steady state of network performance is achieved.
1 . A computer-implemented method of analyzing performance of a network based on simulating transactions through the network by a plurality of instances of an application, wherein the plurality of instances of the application may concurrently run on the network, said method comprising:
initializing respective response times for each instance of the application to an initial value;
generating a series of start times to simulate starting of instances of the application;
simulating simultaneous transactions of the plurality of instances of the application running across the network based on the series of start times;
determining conditions of the network resulting from the simultaneous transactions;
determining respective response times of the instances of the application in the presence of the multiple transactions running across the network until the conditions of the network stabilize; and
providing an output indicating a performance of the network based on the determined response times.
2 . The method of claim 1 , wherein determining conditions of the network comprises determining an average network utilization for duration of the instance of the application.
3 . The method of claim 1 , wherein determining conditions of the network comprises determining a queuing delay.
4 . The method of claim 1 , wherein determining conditions of the network comprises determining an amount of packet loss.
5 . The method of claim 1 , wherein determining the response times of the instances of the application comprises calculating the response times based on a table lookup corresponding to the determined network conditions.
6 . The method of claim 1 , wherein determining the response times of the instances of the application comprises calculating response times based on when the network is over-utilized.
7 . The method of claim 6 , wherein determining the response times includes modeling congestion on a router in the network based on weighted fair queuing (WFQ) properties and the traffic flowing through the router.
8 . The method of claim 6 , wherein determining the response times comprises calculating a probability of packet loss for the transactions running across the network.
9 . The method of claim 1 , further comprising
initializing respective response times for instances of a second application to an initial value;
generating a series of start times to simulate starting of instances of the second application; and
simulating simultaneous transactions of the plurality of instances of the application and second application running across the network based on the series of start times.
10 . The method of claim 9 , further comprising:
determining conditions of the network resulting from the simultaneous transactions; and
determining respective response times of the instances of the application and second application in the presence of the multiple transactions running across the network until the conditions of the network stabilize.
11 . The method of claim 10 , further comprising:
providing an output indicating a performance of the network based on the determined response times for the application and second application.