3 Terminology, Definitions, and Abbreviations 3

PT ××××—××××

Table of Contents

1 Scope 3

2 Normative References 3

3 Terminology, Definitions, and Abbreviations 3

3.1 Terminology 3

3.2 Abbriviations 3

4 Key-Value Based Cloud Data Storage System Technical Architecture 3

5 Functional Requirements of Key-Value Based Cloud Data Storage System 4

5.1 Data Consistency 4

5.2 Overview 4

6 Quality Requirements of Key-Value Based Cloud Data Storage System 5

6.1 Availability 5

6.2 Security 5

PT ××××—××××

Key-Value Based Cloud Storage Interface

1 Scope

The scope of work group is composed of two major parts: Key Value Database and Key Value Cloud Storage.

1) Key Value Database is the first priority work that should be invested. A preliminary KVDB standard draft which based on Key Value Database planned to be released at the end of 2014.

2) Cloud Storage is a complementary choice, as market of Key Value based cloud storage has huge potential. When and how to start KV Cloud Storage work need to be reevaluated along with the growth of cloud storage market and member companies’ support.

The test standard is not included in the scope, as test standard is complex and is hard to accomplish in a very short time. If necessary, a new work group focused on test standard is suggested.

The document specifies the key-value based cloud data storage system technical reference architecture, functional requirements and quality requirements.

The document applies to the design, development, and use of the key-value based cloud data storage system.

2 Normative References

RFC2616 Hypertext Transfer Protocol 1.1 (HTTP/1.1)

3 Terminology, Definitions, and Abbreviations

3.1 Terminology

Key-value based cloud data storage system

3.2 Abbriviations

4 Key-Value Based Cloud Data Storage System Technical Architecture

5 Functional Requirements of Key-Value Based Cloud Data Storage System

5.1 Data Consistency

To ensure the BASE features of data, the cloud data management system requires eventual consistency. Key-value based cloud data management should meet the requirement of strong consistency and can downgrade strong consistency to eventual consistency when availability and fault tolerance predominate.

BASE prioritizes availability and tries to put off the synchronization of system data. All data copies are synchronized, so that the whole system is consistent. The cloud storage Isystem based on the key-value data model should meet the strong consistency requirement, and may downgrade strong consistency to eventual consistency when availability and fault tolerance predominate.

Note: The data management system requires strong consistency or eventual consistency. Strong consistency is required for ACID of relational databases. ACID is in reference to GB/T 17173.1-1997.

5.2 Overview

The application interfaces of key-value-based cloud data management include table-level and row-level operation interfaces.

The table-level operations include the following:

a) List all the tables of a specified account

b) Create tables

c) Delete tables

d) Obtain metadata

e) Add/update metadata

The row-level operations include the following:

a) Insert/update data

b) Delete data

c) Obtain data

d) Query data according to conditional expressions