1

Introduction

Abstract

With the rapid growth in the size of Cloud data, cost-effective data storage has become one of the key issues in Cloud research, yet the reliability of the huge amounts of Cloud data needs to be fully assured. What we investigate in this book is the trade-off of cost-effective data storage and data reliability assurance in the Cloud. This chapter introduces the background knowledge and key issues of this research.

Keywords

background knowledge Cloud storage data reliability assurance technologies data reliability definition Introduction key issue With the rapid growth in the size of Cloud data, cost-effective data storage has become one of the key issues in Cloud research, yet the reliability of the huge amounts of Cloud data needs to be fully assured. In this book, we investigate the trade-off of cost-effective data storage and data reliability assurance in the Cloud. The novel research stands from the Cloud storage service providers' perspective and investigates the issue on how to provide cost-effective data storage service while meeting the data reliability requirement throughout the whole Cloud data life cycle. This topic is important and has a practical value to Cloud computing technology. Especially, for data-intensive applications that are of data-intensive characteristics, our research could dramatically reduce its storage cost while meeting the data reliability requirement and hence has a positive impact on promoting the deployment of the Cloud. This chapter introduces the background knowledge and key issues of this research. It is organized as follows. Section 1.1 gives the definition of data reliability and briefly introduces current data reliability assurance technologies in the Cloud. Section 1.2 introduces the background knowledge related to Cloud storage. Section 1.3 outlines the key issues of the research. Finally, Section 1.4 presents an overview for the book structure.

1.1. Data reliability in the Cloud

The term "reliability" is widely used as an aspect of the service quality provided by hardware, systems, Web services, etc. In Standard TL9000, it is defined as "the ability of an item to perform a required function under stated conditions for a stated time period" [1]. For data reliability specifically, which refers to the reliability provided by the data storage services and systems for the stored data, it can be defined as "the probability of the data surviving in the system for a given period of time" [2]. While the term "data reliability" is sometimes used in the industry as a superset of data availability and various other topics, in this book we will stick to the definition of data reliability given earlier. Data reliability indicates the ability of the storage system to keep data consistent, hence it is always one of the key metrics of a data storage/management system. In large-scale distributed systems, due to the big quantity of storage devices being used, failures of storage devices occur frequently [3]. Therefore, the importance of data reliability is prominent, and these systems need better design and management to cope with frequent failures. Increasing the data redundancy level could be a good way for increasing data reliability [4,5]. Among several major approaches for increasing the data redundancy level, data replication is currently the most popular approach in distributed storage systems. At present, data replication has been widely adopted in many current distributed data storage/management systems in both industry and academia, which include examples such as OceanStore [6], DataGrid [7], Hadoop Distributed File System [8], Google File System [9], Amazon S3 [10], and so forth. In these storage systems, several replicas are created for each piece of data. These replicas are stored in different storage devices, so that the data have better chance to survive when storage device failures occur. In recent years, Cloud computing is emerging as the latest distributed computing paradigm, which provides redundant, inexpensive, and scalable resources in a pay-as-you-go fashion to meet various application requirements [11]. Since the advent of Cloud computing in late 2007 [12], it has fast become one of the most promising distributed solutions in both industry and academia. Nowadays, with the rapid growth of Cloud computing, the size of Cloud storage is expanding at a dramatic speed. It is estimated that by 2015 the data stored in the Cloud will reach 0.8 ZB (i.e., 0.8 × 1021 bytes or 800,000,000 TB), while more data are "touched" by the Cloud within their life cycles [13]. For maintaining such a large amount of Cloud data, data reliability in the Cloud is considered more important than ever before. However, due to the accelerating growth of Cloud data, current replication-based data reliability management has become a bottleneck for the development of Cloud data storage. For example, storage systems such as Amazon S3, Google File System, and Hadoop Distributed File System all adopt similar data replication strategies called the "conventional multi-replica replication strategy," in which a fixed number of replicas (normally three) are stored for all data to ensure the reliability requirement. For storage of the huge amounts of Cloud data, these conventional multi-replica replication strategies consume a lot of storage resources for additional replicas. This could cause negative effects for both the Cloud storage providers and users. On one hand, from the Cloud storage provider's perspective, the excessive consumption of storage resources leads to a big storage overhead and increases the cost for providing the storage service. On the other hand, from the Cloud storage user's perspective, according to the pay-as-you-go pricing model, the excessive storage resource usage will finally be paid by the storage users. For data-intensive Cloud applications specifically, the incurred excessive storage cost could be huge. Therefore, Cloud-based applications have put forward a higher demand for cost-effective management of Cloud storage. While the requirement of data reliability should be met in the first place, data in the Cloud needs to be stored in a highly cost-effective manner.

1.2. Background of Cloud storage

In this section, we briefly introduce the background knowledge of Cloud storage. First, we introduce the distinctive features of Cloud storage systems. Second, we introduce the Cloud data life cycle.

1.2.1. Distinctive features of Cloud storage systems

Data reliability is closely related to the structure of the storage system and how the storage system is being used. Different from other distributed storage systems, the Cloud storage system has some distinctive features that could either be advantages or challenges for the data reliability management of Cloud data.

1.2.1.1. On-demand self-service and pay-as-you-go pricing model

The on-demand usage of Cloud storage service and pay-as-you-go payment fashion have greatly facilitated the storage users that they only need to pay for the resources used for storing their data for a needed time period. The cost is easy to be estimated according to the size of data generated [14]. However, based on the pay-as-you-go model, every usage of the resources can be strictly reflected onto the bills payable at the end of the month. Therefore, minimizing resource consumption becomes demanding and critical. This principle is not only applicable to the service users, but also to the Cloud storage service providers. In most current Cloud storage services, excessive data redundancy is compulsorily generated to ensure data reliability. For data-intensive applications, such excessive data redundancy consumes a large amount of storage resources, and hence incurs a very high cost.

1.2.1.2. Redundant and scalable virtualized resources

In the Cloud, large amounts of virtualized computing and storage resources are pooled to serve users with various demands [1]. Redundant computing resources make it easy to conduct parallel processing, while the redundant storage resources make it easy to distribute data. For meeting a higher computing/storage demand, the resource pool can be scaled out rapidly, and the virtualization keeps the complex procedures transparent from the service users. However, the virtualization of resources has also led to a challenge that various kinds of data reliability requirement need to be fully assured to make the Cloud storage service trustworthy.

1.2.1.3. Dedicated Cloud network

Cloud systems (public Clouds specifically) are primarily running based on data centers with dedicated networks, which interconnect with each other using dedicated links [15]. Such a dedicated feature of the Cloud network has provided the Cloud the potential of full bandwidth control ability. The Cloud storage system could benefit from the dedicated Cloud network that the creation and recovery of data can be conducted in a fully controllable and predictable manner. At the meantime, there is still a great potential that data transfer in the Cloud network could be optimized for being conducted more cost-effectively.

1.2.1.4. Big data

"Big data" is the term for a collection of data sets so large and complex that it becomes...