Load Balancing Method and Apparatus, and Software Streaming System Using the Same

• US 20090125625A1
• Filed: 07/27/2006
• Published: 05/14/2009
• Est. Priority Date: 09/15/2005
• Status: Active Grant

• Current Assignee: Electronics and Telecommunications Research Institute
• Original Assignee: Electronics and Telecommunications Research Institute
• Inventors: Hur, Sung Jin, Choi, Wan, Kang, Sung Joo, Shim, Jeong-Min
• Granted Patent: US 7,788,380 B2
• Time in Patent Office: Days
• US Class Current: 709/226
• CPC Class Codes: H04L 67/1001 for accessing one among a plurality of replicated servers
• CPC Class Codes: H04L 67/1008 based on parameters of servers, e.g. available memory or workload monitoring of computer activity G06F11/30

Abstract

Disclosed herein is a load balancing method and apparatus which can efficiently distribute load to a plurality of servers providing a software streaming service where a bandwidth-on-demand is different depending on sending times, and a software streaming system using the same. Sessions established between the servers and clients are classified into first sessions which send data for initial execution and second sessions which send data desired by users after the initial execution is completed, and the resulting session information of each of the servers is collected. A total load amount of each of the servers is calculated under the condition that different load weights are applied to the first and second sessions. A server with a minimum one of the total load amounts is selected from among the servers by comparing the total load amounts of the servers with one another, and the selected server is allocated to a new client requesting the software streaming service.

Claims

1. A load balancing method for distributing load to a plurality of servers which provide a software streaming service to clients, comprising:

classifying sessions established between the servers and the clients into first sessions which send data for initial execution and second sessions which send data desired by users after the initial execution is completed, and collecting the resulting session information of each of the servers;

calculating a total load amount of each of the servers under the condition of applying different load weights to the first and second sessions; and

comparing the total load amounts of the servers with one another to select a server with a minimum one of the total load amounts from among the servers when the software streaming service is requested by a new client, and allocating the selected server to the new client such that the selected server provides the software streaming service to the new client.

2. The load balancing method as set forth in claim 1, wherein collecting the resulting session information of each of the servers includes collecting a software streaming service request signal from each of the clients, and an initial execution completion signal and a service termination signal from each of the servers, the initial execution completion signal

indicating completion of the sending of the initial execution data, the service termination signal indicating session release.

3. The load balancing method as set forth in claim 2, wherein collecting the resulting session information of each of the servers further includes:

incrementing each of the number of total sessions and the number of the first sessions of the server allocated to the new client by one if the software streaming service request signal from the new client is received;

decrementing the number of the first sessions of an arbitrary one of the servers by one if the initial execution completion signal from the arbitrary server is received; and

decrementing the number of total sessions of the arbitrary server by one if the service termination signal from the arbitrary server is received.

4. The load balancing method as set forth in claim 3, wherein calculating the total load amount includes calculating the total load amount of each of the servers using the following equation:

T=I×

b+{(S−

I)×

a}

where T is the total load amount, I is the number of the first sessions, S is the number of total sessions, b is the load weight applied to the first sessions, and a is the load weight applied to the second sessions.

5. The load balancing method as set forth in claim 1, wherein a ratio between the load weights of the first and second sessions is set in proportion to a ratio between bandwidths-on-demand in the first and second sessions.

15. The load balancing method as set forth in claim 4, wherein a ratio between the load weights of the first and second sessions is set in proportion to a ratio between bandwidths-on-demand in the first and second sessions.

6. A load balancing apparatus for distributing load to a plurality of servers which provide a software streaming service to clients, comprising:

a transmitter for transmitting information of an allocated one of the servers to a new client requesting the software streaming service;

a receiver for receiving load information from the clients and servers;

a load information table for managing information about sessions established between the servers and the clients on a server-by-server basis under the condition of classifying the sessions into first sessions which send initial execution data and second sessions which send data desired by users after the sending of the initial execution data is completed; and

a controller for changing values of the load information table on the basis of the load information received by the receiver and allocating one of the servers having a minimum load amount to the new client requesting the software streaming service on the basis of information recorded in the load information table.

7. The load balancing apparatus as set forth in claim 6, wherein the load information table includes:

the number of total sessions of each of the servers;

the number of the first sessions among the total session number;

load weights applied respectively to the first and second sessions; and

a total load amount of each of the servers calculated by applying the load weights to the first and second sessions.

8. The load balancing apparatus as set forth in claim 7, wherein the controller increments each of the number of the total sessions and the number of the first sessions of the server allocated to the new client in the load information table by one if a software streaming service request signal from the new client is received, decrements the number of the first sessions of an arbitrary one of the servers in the load information table by one if an initial execution completion signal from the arbitrary server is received, and decrements the number of the total sessions of the arbitrary server in the load information table by one if a service termination signal from the arbitrary server is received.

9. The load balancing apparatus as set forth in claim 7, wherein the total load amount of each of the servers in the load information table is calculated using the following equation:

T=I×

b+{(S−

I)×

a}

where T is the total load amount, I is the number of the first sessions, S is the number of the total sessions, b is the load weight applied to the first sessions, and a is the load weight applied to the second sessions.

10. The load balancing apparatus as set forth in claim 7, wherein a ratio between the load weights of the first and second sessions in the load information table is set in proportion to a ratio between bandwidths-on-demand in the first and second sessions.

11. A software streaming system comprising:

a plurality of servers for providing a software streaming service;

one or more clients for receiving the software streaming service; and

a load balancer for managing information about sessions established between the servers and the clients on a server-by-server basis under the condition of classifying the sessions into first sessions which send data for initial execution and second sessions which send data desired by users after the initial execution is completed, and allocating one of the servers having a minimum load amount to a new client requesting the software streaming service, on the basis of the session information,

where in the new client sends a software streaming service request signal to the load balancer in response to a request from an associated user, and accesses the server allocated by the load balancer to receive the software streaming service therefrom.

12. The software streaming system as set forth in claim 11, wherein the load balancer includes a load information table, the load information table including the number of total sessions of each of the servers, the number of the first sessions among the total session number, load weights applied respectively to the first and second sessions, and a total load amount of each of the servers calculated by applying the load weights to the first and second sessions.

13. The software streaming system as set forth in claim 12, wherein:

each of the servers provides an initial execution completion signal and a service termination signal to the load balancer, the initial execution completion signal indicating completion of the sending of the initial execution data, the service termination signal indicating session release; and

the load balancer increments each of the number of the total sessions and the number of the first sessions of the server allocated to the new client in the load information table by one if the software streaming service request signal from the new client is received, decrements the number of the first sessions of an arbitrary one of the servers in the load information table by one if the initial execution completion signal from the arbitrary server is received, and decrements the number of the total sessions of the arbitrary server in the load information table by one if the service termination signal from the arbitrary server is received.

14. The software streaming system as set forth in claim 12, wherein the load balancer calculates the total load amount of each of the servers using the following equation:

T=I×

b+{(S−

I)×

a}

Where T is the total load amount, I is the number of the first sessions, S is the number of the total sessions, b is the load weight applied to the first sessions, and a is the load weight applied to the second sessions.