KR20050021087A - Cache consistency maintaining method on wireless internet circumstance - Google Patents

Abstract

PURPOSE: A method for keeping cache consistency on a wireless Internet environment with adjustment of a lease period is provided to effectively adjust workload of a server and a network by shortening or lengthening the lease period of a specified cache object in association with an access frequency of a client. CONSTITUTION: The client requests the specified cache object to the server. The server updates access frequency data for access of entire server by recording access to the cache object from the client as a number. The lease periodically for the cache object is calculated by referring to an access frequency data. The requested cache object is provided to the client and the lease period is determined.

Description

CACHE CONSISTENCY MAINTAINING METHOD ON WIRELESS INTERNET CIRCUMSTANCE}

The present invention relates to a method of maintaining cache coherency in a wireless Internet environment, and more particularly, to a method of maintaining cache coherency in a wireless Internet environment in which a workload can be distributed between a server and a client by adjusting a lease period.

One of the important areas of research in data caching is cache coherency, which ensures that the data currently stored in the cache matches the original data held by the server.

One of the cache coherency techniques that guarantees that users never provide out-of-date data is called strong cache coherence technique.

The lease algorithm is one of the most recent strong cache coherence techniques and is known to have a very effective effect of properly distributing server and client loads.

The lease algorithm may be supplemented and applied to a wireless environment having frequent disconnection of clients. In this case, the lease period is an important performance factor that can control the degree of workload distribution between the server and the client.

The lease algorithm introduces the concept of a lease, which is a time-related variable, so that when a client requests data from the server, it is assigned a lease first, and every time the server updates its data during the lease period after the client takes the data. An invalidation message is sent to the client to indicate that the data has been updated.

After the lease period, the client reclaims the lease from the server as needed to maintain cache consistency. As a result, the server side only needs to manage a valid client for the lease, thereby reducing the workload. On the client side, the valid lease among the cache data owned by the client can immediately respond to the user's request, thereby reducing the response time. It can be improved.

However, in the prior art as described above, in a wireless environment, there is another aspect of maintaining cache coherence, and the lease algorithm can be efficiently applied to the wireless environment through this consideration. At this time, the longer the lease period, the more the status information of the client that the server needs to maintain increases the workload of the server, and the shorter the lease period, the greater the frequency of invalidation messages sent to the client, which places a heavy load on the network. . In other words, depending on the lease period, the workload of the server and the workload of the network are in conflict, and a criterion for adjusting the lease is required.

Accordingly, the present invention has been made in view of the above necessity, and the wireless Internet can effectively control the workload of servers and networks by lengthening or shortening the lease period of a particular cache object in association with the client's connection frequency. The goal is to provide a way to maintain cache coherency in the environment.

The present invention for achieving the above object comprises the steps of requesting a specific cache object from the client to the server; Updating the access frequency data for the total number of connections of the server by recording the number of accesses of the client in the server; Calculating a lease period for the cache object with reference to the access frequency data; And providing a cache object requested to the client and determining a lease period.

Hereinafter, exemplary embodiments of the present invention will be described in detail with reference to the accompanying drawings.

In the present invention, particularly in an information inquiry service using a mobile terminal, a particular cache object having a high frequency of service access by the mobile terminal is selected to set a long lease period for these objects, and a short lease period is set for a cache object having a low access frequency. We propose a method to efficiently manage server workload.

The information subjects of the information retrieval service are server and client.

The client is also provided with a specific cache object requested from the server and given a lease duration. During the lease period, the client receives an invalidation message for the cache object from the server and checks whether the cache object is invalidated after the lease period.

The server has data that records the frequency of access to individual cache objects, and when there is a request for a cache object from a client, the server determines the lease period by referring to the data of access frequency.

To determine the lease duration, the server updates the frequency record frequently and describes its behavior.

When a request for a cache object is made from a client, the server increases the number of accesses to the cache object. Then, the number of connections to individual cache objects of the entire server is calculated. The total number of connections in the server is the sum of the number of accesses of individual cache objects.

The server calculates the access frequency by calculating the ratio of the total number of accesses to the server and the number of accesses. Increasing the number of accesses increases the value of the access frequency, and conversely decreases the access frequency value.

The lease duration is calculated by multiplying the frequency of access by the maximum lease duration. Here, it can be seen that the sum of the individual lease periods results in the maximum lease period. Therefore, the maximum lease period is divided in proportion to the access frequency and allocated as the lease period of the individual cache object.

1 is an operation flowchart of a cache coherence maintaining method in a wireless Internet environment according to an embodiment of the present invention, which includes requesting a specific cache object from a client to a server as shown in FIG. Increasing the number of accesses to the cache object requested by the server; Calculating an access frequency by calculating a ratio of the total number of accesses to the server and the number of accesses; Calculating a lease period by multiplying the access frequency by the maximum lease period; Providing the requested cache object to the client and determining a lease duration.

Assume that a client accesses specific content using an information retrieval service.

The client requests a specific cache object from the information inquiry service to the server.

The server prepares a service for the cache object by requesting the cache object from the client, and determines the lease period of the cache object.

To determine the lease duration, the server determines the lease duration based on this access frequency by referring to the data recording the frequency of the client's access to the cache object. Prior to this, the server records the number of accesses to the requested cache object in the access frequency record data.

Then, the access frequency is calculated by calculating the ratio of the total number of connections of the server and the number of accesses. The lease duration is calculated by multiplying the access frequency by the maximum lease duration.

The server provides the requested cache object to the client and determines the lease duration.

In summary, the lease period is prolonged in proportion to the cache object with high access frequency, and the lease period is shortened with the cache object with low access frequency so that each client can check whether or not to invalidate.

As described in detail above, the present invention determines the lease period based on the frequency of the client's access to a particular cache object when determining the lease period, which is one of the important performance factors that determine the workload of the server and the network. This has the effect of promoting efficient workload distribution.

1 is a flowchart illustrating a cache coherency maintaining method in a wireless Internet environment according to an exemplary embodiment of the present invention.

Claims (3)

Requesting a specific cache object from the client to the server; Updating the access frequency data for the total number of accesses of the server by recording accesses to the cache object of the client by the number of times by the server; Calculating a lease period for the cache object with reference to the access frequency data; And providing a cache object requested to the client and determining a lease period. The method of claim 1, wherein updating the access frequency data comprises: increasing the number of accesses to the cache object requested at the server; And calculating an access frequency by calculating a ratio of the number of accesses to the cache objects of the entire server and the number of accesses. 2. The method of claim 1, wherein the lease period is calculated by multiplying the maximum frequency and the maximum frequency recorded in the access frequency data.