Browse Prior Art Database

A system and method to create a cost efficient mobility solution of storing data in public clouds Disclosure Number: IPCOM000232553D
Publication Date: 2013-Nov-15
Document File: 9 page(s) / 186K

Publishing Venue

The Prior Art Database


This disclosure discloses method of buidling a cost efficient data mobility solution, of storing data in public clouds. The solution includes modules of doing data character and access behivior collection , statistic, and prediction. Based on those information, in conjunction with pricing model of diffrent cloud storage vendors, lowest price of storing certain data could be calculated. By migrating data to cloud storage with lowest price, the overall cost of storing the data during its full lifecycle could be reduced.

This text was extracted from a PDF file.
This is the abbreviated version, containing approximately 38% of the total text.

Page 01 of 9

A system and method to create a cost efficient mobility solution of storing data in public clouds

Cost efficiency on public cloud

Although the principle of cloud storage is providing data storage in lower TCO(Total Cost of Ownership), different cloud vendor provides different pricing model. For example, some of the cloud storage package charges more on data store space basis but less on data transfer, and other cloud storage doing vice versa. But certain data has different access pattern during its full life cycle, lock-in to certain pricing model of single vendor will impact the cost efficiency and flexibility of utilizing public cloud storage.

For example, there are two public cloud storage services, and they have different pricing model for first 1TB:

Cloud Service Vendor A: store $0.095/GB/month, $0.005/1000 requests (put, post, copy, list), outgoing data rate $0.12/GB;

Cloud Service vendor B: store $0.1/GB/month, free for all requests, outgoing data rate $0.12/GB;

As an easy model, if the data amount is 1GB, there are more than 1000 requests happened during a month, it's cheaper to store onto Cloud service vendor B.

So there are several problems to make decision to store data onto public cloud:

Each public cloud has different pricing model based on the capacity, request, and out going data rate, it has tiered offering;

It's hard for people to figure out the access pattern and figure out which public cloud is the most suitable one for his access pattern;

This disclosure provides a new method to the data placed in the public storage cloud vendor from theenterprise customer

The key innovation points of the solution are ,

1. Access pattern of a file will be tracked and stored on metadata server.

2. Based on the statistic, the system will make prediction of size and data transfer rate in a certain period for this file.

3. With this prediction, the cost of storing the file could be calculated. And the system will store the file to cloud service with lowest cost.

4. Also according to behaviors of files during its life cycle, the data could be move between clouds with different cost model, for the purpose of saving cost.

The overall architecture is depicted in the figure below .


Page 02 of 9

The key advantages of this method :


Page 03 of 9

1. Maximize cost efficiency and flexibility of storing data on public cloud , avoid cost impact due to single vendor lock -in.

2. Combination of behavior analysis based pre -transfer prediction and migration of existing data

3.0 Overall Solution

Here is the overall component diagram of the solution:

In the applications side, Data Access Reporter will record the data operation behavior and report to metadata management server;

In the metadata management server side:

Data Access Collector will collect all the behavior from front end applications;

Pattern Predictor will calculate out the behavior pattern for files;

Data Store Manager will utilize the predicted pattern and send out guidance to fro...