System and method for performing capacity planning for enterprise applications

• US 20060013134A1
• Filed: 06/28/2005
• Published: 01/19/2006
• Est. Priority Date: 06/28/2004
• Status: Active Grant

• Current Assignee: HyPerformix, Inc. (Broadcom, Inc.)
• Original Assignee: HyPerformix, Inc. (Broadcom, Inc.)
• Inventors: Neuse, Douglas M.
• Granted Patent: US 8,200,805 B2
• Time in Patent Office: Days
• US Class Current: 370/230
• CPC Class Codes: H04L 41/082 the condition being updates or upgrades of network functionality
• CPC Class Codes: H04L 41/142 using statistical or mathematical methods
• CPC Class Codes: H04L 41/147 for predicting network behaviour
• CPC Class Codes: H04L 43/0882 Utilisation of link capacity
• CPC Class Codes: H04L 43/0894 Packet rate

Abstract

A system and method for capacity planning for enterprise networks, such as identifying bottlenecks and removing or replacing the bottleneck device are provided. The device utilization for one or more network devices are measured or read from measured data. A relative load is calculated from the device utilization data and device utilization is compared to a device threshold to determine the bottleneck device. A method is also provided for determining network utilizations, network populations and a relative response times based on only limited measurable device usage data.

Claims

1. A computer hardware system executing software programs for performing network capacity planning and capacity upgrading of a network into a new configuration, the system comprising:

a means for reading device utilization data for a network device;

a means for storing the device utilization data;

a means for calculating a utilization ratio from the device utilization data; and

a means for predicting a bottleneck device by analyzing the utilization ratio.

2. The system of claim 1 wherein the utilization ratio is calculated from a predefined utilization.

3. The system of claim 1 further comprising

a software application operating on the network;

a means for calculating network response times for said software application.

4. The system of claim 1 further comprising a means for predicting a plurality of bottleneck network devices.

5. The system of claim 1 further comprising a means for upgrading the bottleneck device of the network comprising:

a means for identifying a new device with a faster benchmark rating than the bottleneck device; and

a means for predicting device utilization with the new device in the new network configuration.

6. The system of claim 4 further comprising a means for upgrading a plurality of bottleneck devices of the network comprising:

a means for identifying new devices with a higher benchmark rating than the bottleneck devices; and

a means for predicting relative load device utilizations with the new devices in the new network configuration.

7. A method for performing capacity planning for a network, said network having a set of network devices, each network device having a device utilization threshold, comprising the steps of:

providing measured usage data for the set of network devices which includes device utilizations for a subset of the set of network devices;

identifying a bottleneck device from the set of network devices using the measured network usage data.

8. The method of claim 7 further comprising the step of computing unmeasured usage data from the measured usage data.

9. The method of claim 8 further comprising the step of identifying a bottleneck device from the one or more network devices using the measured usage data and the unmeasured usage data

10. The method of claim 7 further comprising the step of removing the bottleneck device.

11. The method of claim 7 further comprising the step of replacing the bottleneck device with an updated network device of a faster benchmark rating.

12. The method of claim 7 wherein the step of providing measured usage data for the one or more network devices comprises the step of reading from monitored network performance data the measured usage data.

13. The method of claim 7 wherein the step of identifying a bottleneck device further comprises the step of forming a utilization ratio of device utilization threshold to device utilization for each network device in the set of network devices.

14. The method of claim 13 wherein the step of identifying a bottleneck device further comprises sorting the utilization ratios.

15. The method of claim 13 wherein the step of identifying a bottleneck device further comprises finding the minimum utilization ratio.

16. The method of claim 8 wherein the step of computing unmeasured usage data from the measured usage data comprises the further step of calculating network response times

17. The method of claim 8 wherein computing unmeasured usage data from the measured usage data comprises the further step of predicting device utilization.

18. The method of claim 8 wherein the step of computing unmeasured usage data from the measured usage data comprises the further step of predicting network population.

19. A method for determining capacity parameters for a network including a set of network devices and operating with a given base load comprising:

providing a measured device utilization at the base load;

calculating a network arrival rate from the measured device utilization;

calculating device utilization at a relative load value; and

calculating relative load network population at a relative load value.

20. The method of claim 19 wherein the step of providing a measured device utilization at a base load comprises acquiring data from a system performance monitor.

21. The method of claim 19 wherein the step of calculating a network arrival rate from the measured device utilization at a base load comprises setting the base load value to a constant value.

22. The method of claim 19 that further comprises the step of calculating a network response time at a relative load value.

23. A method for performing network capacity planning on a network having a set of network devices and having limited measurable network device usage data available, comprising the steps of:

identifying a subset of busiest network devices from the set of network devices.

measuring device utilizations for the subset of busiest network devices; and

estimating a network population ratio increase from the measured device utilizations;

24. The method of claim 23 comprising the further step of calculating a relative response time based on the network population ratio increase.

25. The method of claim 23 wherein the step of identifying a subset of busiest network devices from the set of network devices comprises the additional step of monitoring network traffic from the set of network devices.

26. The method of claim 23 further comprising the step of verifying the accuracy of the step of estimating the network population ratio.

27. A method for performing a capacity upgrade on a given network having a set of network devices to an upgraded network comprising the steps of:

normalizing stored network device speeds associated with network devices in the given network to a constant value;

replacing one or more network devices in the given network with one or more new network devices;

calculating relative speeds for the one or more new network devices;

predicting network device utilizations for network devices in the upgraded network for one or more relative load values;

28. The method of claim 27 in which the step of replacing one or more network devices is performed after the step of calculating relative speeds for the new network devices

29. The method of claim 27 in which the step of replacing one or more network devices is performed before the step of normalizing stored network device speeds.

30. The method of claim 27 wherein the step of predicting device utilizations at a relative load value comprises the step of calculating device utilizations at a base load.

31. A method for determining network response times within the process of performing network capacity planning for a given network, the method comprising the steps of:

obtaining network usage data; and

calculating relative network response times at one or more relative load values based on the network usage data.

32. The method of claim 31 wherein obtaining the network usage data includes the act of measuring network usage data.

33. The method of claim 31 wherein the step of obtaining the network usage data includes the measurement of an absolute network response time.

34. The method of claim 31 wherein the step of obtaining the network usage data includes the step of calculating relative load network population from prior measured usage data.

35. The method of claim 31 wherein calculating relative network response times at one or more relative load values further comprises:

defining a base load that is associated with the given network;

normalizing the relative network response time at the base load to a constant value;

calculating a the relative response time R(x) for the relative load values according to the formula;

R(x)=(N₀(x)/ N₀(1))/x

wherein x is a relative load value;

N₀(x) is the relative load network population at relative load value x and N₀(1) is the base load network population at the base load x=1.

36. The method of claim 31 further comprising calculating absolute network response times corresponding to relative network response times at one or more relative load values wherein calculating said absolute response time includes multiplying a relative response time at a relative load value by the absolute response time at the base load.

37. The method of claim 34 wherein the constant value is set equal to one in the step of normalizing the relative response time at the base load.

38. The method of claim 31 wherein the step of calculating network response times at one or more relative load values comprises the further steps of:

fitting response times to a known functional form, the relative load value being the dependent variable in said functional form;

storing the constants generated by the functional fit in a computer memory;

utilizing the functional fit in the capacity planning process to predict network response times for the relative load values.

39. The method of claim 38 wherein the response time functional fit is a function of the form

R(x)=α

₀+α

₁x+α

₂(N₀(x)/N₀(1))/x

wherein α

₀, α

₁, and α

₂ are constants, x is the relative load value, N₀(x) is the relative load network population at relative load value x, and N₀(1) is the relative load network population at the base load x=1.