Efficient Data Center Monitoring

• US 20150085695A1
• Filed: 09/22/2014
• Published: 03/26/2015
• Est. Priority Date: 09/20/2013
• Status: Abandoned Application

• Current Assignee: New Relic
• Original Assignee: CoScale NV (New Relic)
• Inventors: RYCKBOSCH, Frederick, POLFLIET, Stijn
• Application Number:
  US14/492,176
• Publication Number:
  US 20150085695A1
• Time in Patent Office: Days
• US Class Current: 370/253
• CPC Class Codes: H04L 43/02 Capturing of monitoring data
• CPC Class Codes: H04L 43/028 by filtering
• CPC Class Codes: H04L 43/04 Processing captured monitoring data, e.g. for logfile generation
• CPC Class Codes: H04L 43/045 for graphical visualisation of monitoring data
• CPC Class Codes: H04L 43/0876 Network utilisation, e.g. volume of load or congestion level
• CPC Class Codes: H04L 43/0894 Packet rate
• CPC Class Codes: H04L 43/12 Network monitoring probes
• CPC Class Codes: H04L 47/196 Integration of transport layer protocols, e.g. TCP and UDP

Abstract

A system for monitoring network traffic and resource usage in a data center. The system comprises a node comprising an input terminal for capturing network packets. The node comprises a processor for running a packet analyzer for analyzing the network packets classifying packets in request/response pairs per type based on deep packet inspection and storing the request/response pair types together with time stamps. The node comprises a memory configured for storing the request/response pairs. The system comprises a collector module configured for collecting the request/response pairs. The system comprises at least one server running a simple agent programmed for transferring resource usage. The system comprises a correlation module programmed for correlating the request/response pairs information with the resource usage.

Claims

1. A system for monitoring network traffic and resource usage in a data center, the system comprising:

a node comprising an input terminal configured for capturing network packets, a processor programmed for running a packet analyzer for analyzing the network packets whereby packets are classified in request/response pairs per type based on deep packet inspection, and a memory configured for storing the request/response pairs,

a collector module configured for collecting the request/response pairs and for receiving resource usage information from at least one server, and a correlation module programmed for correlating the request/response pairs information with the resource usage.

2. A system according to claim 1, whereby the input terminal captures substantially all network packets of the network traffic and whereby said substantially all network packets are processed by the packet analyzer.

3. A system according to claim 1 whereby the network packets first are filtered based on packet information and the packet analyzer only processes the filtered packets.

4. A system according to claim 3, whereby only the HTTP packets are passing through the filter, filtering on the TCP port data, and whereby the packet analyzer parses URL and host information from the filtered packets.

5. A system according to claim 1 whereby response times of several request of the same type are combined in a latency histogram.

6. A system according to claim 1 whereby the packet analyzer processes the packets in order of arrival and/or whereby the packet analyzer processes the packets classifying them according to user defined request types.

7. A system according to claim 1, whereby data in the packets is aggregated and/or whereby the packet analyzer stores only 10-100 requests per second per request type completely for retrieving full URL and host information.

8. A system according to claim 1, the system furthermore comprising a monitoring unit configured for displaying a data center load versus request/response pair rate.

9. A system according to claim 1, the system being implemented as a computer or the system being implemented as an application.

10. A system according to claim 1, wherein the correlation module is programmed for determining the relationship between the number of requests and the resource utilization.

11. A system according to claim 1, wherein the correlation module is programmed for determining information indicative of how many resources each type of request/response pair uses.

12. A system according to claim 1 wherein the request/response pairs relate to features of the application or defined criteria about an application and wherein the correlation module provides correlation information between the network traffic and the resource usage at an intra-application message level.

13. A method for monitoring network traffic and resource usage in a data center, the method comprising the following steps:

capturing network packets entering or leaving the data center by an input terminal of a node,

analyzing of the network packets by a packet analyzer running on a processor whereby the network packets are classified in request/response pairs per type and this based on deep packet inspection,

obtaining resource usage information of at least one server,

correlating the request/response pairs information with the resource usage.

14. A method according to claim 13 whereby in the capturing step all network packets are captured, and whereby in the analyzing step all captured network packets are analyzed.

15. A method according to claim 13 whereby the captured network packets are filtered based on packet information and whereby the packet analyzer only processes the filtered packets.

16. A method according to claim 13, whereby only the HTTP packets are passing through the filter in the filtering step, filtering on the TCP port data, and whereby the packet analyzer parses URL and host information from the filtered packets in the analysis step and/or wherein the method includes combining response times of several requests in a latency histogram.

17. A method according to claim 13 in at least some of the steps the packets are processed in order of arrival.

18. A data carrier comprising a set of instructions for, when executed on a computer, monitoring network traffic and resource usage in a data center, the monitoring comprising capturing network packets entering or leaving the data center by an input terminal of a node,

analyzing of the network packets by a packet analyzer running on a processor whereby the network packets are classified in request/response pairs per type and this based on deep packet inspection,

obtaining resource usage information of at least one server,

correlating the request/response pairs information with the resource usage.