Basic Introduction

BOS provides four storage types: standard storage, infrequency storage, cold storage and archive storage, to provide users with storage options at different costs. Industry products also usually provide "lifecycle" function. After configuration simply by users, the system will automatically sink the data to reduce the storage cost of users. However, users usually have a problem. How to configure the lifecycle of data? If the number of days allocated is not enough, the hot data will sink to cold data too early, which will make it inconvenient to access the business and incur unnecessary retrieval costs during the retrieval. If the number of days is too much, the gradually cooled data that is not often used on the business side will still be stored at a higher cost, resulting in higher storage expenses.

Based on the above considerations, Baidu AI Cloud BOS has launched the "Intelligent Business Analysis" function. Through in-depth analysis based on users' usage of data, it scientifically predicts the continuous usage frequency of data by subsequent customer businesses, thus giving users the most economical lifecycle configuration strategy.

Instructions Before Use

BOS currently has four storage types: The capacity prices of standard storage, infrequency storage, cold storage and archive storage decrease in turn, but infrequency storage, cold storage and archive storage need to pay corresponding retrieval fees when retrieving. In addition, the performance of the four storage types is also different. For details, please refer to Hierarchical Storage.

BOS intelligent service analysis will help you scientifically analyze the best number of days for standard storage to settle to infrequency storage, thus helping you to pay the minimum storage capacity cost + data retrieval cost under the same service access.

Note: BOS Intelligent Business Analysis will not suggest to you the number of days of settlement from standard storage or infrequency storage to cold storage for the time being, because cold storage is mainly used in cold standby scenarios, and its performance will be quite different from that of the first two storage types. If you rashly settle hot data into cold storage too early, it may affect your performance experience when accessing on the business side.

Operation Steps

1.Log in the BOS console, find the intelligent business analysis function entry in the global overview page, and click open service.

2.After opening the service, you need to define the file directory you expect to conduct intelligent analysis

You can use bucket as the analysis path directly, and all files in the bucket will be used as samples for analysis; you can also analyze some directory prefixes in the bucket.

Note:

• When the configuration is submitted for the first time, or a new path configuration to be analyzed is added later, the system will analyze the user's access behavior to the files under the path from the date of submitting the configuration. The system will not analyze the historical access behavior before the directory.
• For the newly configured path, the longer the system monitoring time is, the more accurate the judgment will be. Therefore, the system will provide you with the first analysis report in 3 days based on the newly submitted configuration.

3.The report generated by the system is shown in the figure below

Key information in the figure:

• Abscissa: Object Age, is file life in days. It can be understood as the time from the upload time to the present.
• Ordinate: It refers to the user’s access to each age range file in GB.
• Purple line: It represents the actual usage of the user. Taking a point in the purple line x = 30, y = 400 as an example, that is to say, all files in the directory have a life span of 30 days. On average, 400GB will be used by users every day.
• Blue line: It is a baseline, representing the critical value of transmission between standard storage and infrequency storage in combination with online price, that is, the consumption amount of standard or infrequency storage is the same in this trend of access behavior.

• Little red flag: That is, the intersection of blue line and purple line represents that under this object age, user configuration lifecycle settlement is the most cost-effective.

  4.Interpretation of policy suggestions:

Based on the trend relationship between the blue line and the purple line, BOS will give intelligent policy suggestions based on various situations, that is, the green text part at the bottom of the figure above. Because there are many trends of blue line and purple line, the policies given by BOS system are as follows:

• When the analysis shows that the access to this part of files is always frequent, BOS will recommend users to use the standard storage type all the time instead of the lifecycle settlement policy.
• When the analysis shows that the access to this part of files has been less, BOS will recommend that users directly configure this part of files for infrequency storage.
• When the analysis shows that the access behavior of this part of files has rules to follow, with more hot data access and less cold data access, BOS will recommend configuring the lifecycle settlement policy at the critical point.
• When the analysis shows that the access rule of this part of files is weak, BOS will also give corresponding settlement days suggestions based on the analysis data. As shown in above figure, through the analysis of BOS system, it is recommended that the user configure the lifecycle policy of standard storage settling to infrequency storage for 36 days.

At present, the analysis function of BOS intelligent business is in hot public beta, you can click to Open Beta Test Application.