Cloud Storage Technology Handbook

Cloud storage products can be said to have thousands of species, can be said to be few, depending on the extent of the definition of cloud storage or how narrow. Regardless of the scope of the definition, the number of cloud storage products is expected to increase this year, because the cloud storage has maintained a rapid momentum of development.

The definition of cloud storage

The problem with cloud storage is that there are so many in its infancy that there are divergent interpretations of the term. To make the definition of cloud storage clearer, industry analysts who often deal with corporate IT experts rearrange the definition of cloud storage, outlining some of the characteristics of cloud storage.

Cloud storage products can be said to have thousands of species, can be said to be few, depending on the extent of the definition of cloud storage or how narrow. Regardless of the scope of the definition, the number of cloud storage products is expected to increase this year, because the cloud storage has maintained a rapid momentum of development.

"The problem with cloud storage is that there are so many in its infancy that there are divergent interpretations of the term." "says George Crump, founder of Swiss company Storage, a consultancy firm based in Texas.

To make the definition of cloud storage clearer, industry analysts who often deal with enterprise IT experts rearrange the definition of cloud storage, outlining the following characteristics of cloud storage:

Ultra-strong scalability

Not subject to specific geographical location

Based on business components

Charge 15 cents per g.

Can be used across different applications

Software as a service (SAAS) product, such as Salesforce.com and Google Apps, does not meet the conditions of cloud storage, although they provide the user with the ability to save data. "If you're a developer, you can't write to their storage infrastructure." So, we think that's not a storage tool. Said Adam Couture, a Gartner analyst.

Couture that there are only a few providers in the cloud storage industry: Amazon.com,nirvanix and Rackspace hosting Mosso cloud. However, he says the list will grow with some companies introducing new products, including at least three well-known companies that plan to launch a S3 (simple Storage Service) similar to Amazon.

Cloud storage products and applications

Cloud storage products can be said to have thousands of species, can be said to be few, depending on the extent of the definition of cloud storage or how narrow. Cloud storage is not a good choice for a transaction-processed database or temporary storage. However, cloud storage is significant for businesses that have unpredictable storage requirements, need low-cost storage arrays, or low-cost long-term archiving.

Cloud storage is beginning to invade the "turf" of cloud services, and some online data backup services are also involved in cloud services, such as Carbonite's online backup, EMC's Mozy, Iron, according to a IDC analyst, Rick Villars. Kings Company's Digital, Seagate Marvell Company's Evault and Symantec Company's online Backup.

In addition, this week IBM launched a service that has not yet been named. This service provides users with its Tivoli continuous data protection product, which backs up information to private cloud storage. By the end of March, the service will also combine the application of IBM's acquisition of Digital Arsenal FX a technology acquired.

"So, is mozy a cloud store?" Mozy is a very advanced service, backup service, with a very good architecture. Its application itself utilizes a cloud storage device. "Villars said. "Mozy is a service built on cloud storage. ”

According to Villars, such a cloud service is inherently not called cloud storage, and storage is only a by-product of another service.

As for Mozy itself, it does not position itself as a cloud storage. "It is not stored as a service, it is backed up as a service." "Decho Company PR director Devin said.

Cloud storage is not a good choice for a transaction-processed database or temporary storage. However, cloud storage is significant for businesses that have unpredictable storage requirements, need low-cost storage arrays, or low-cost long-term archiving.

In particular, companies such as Behance, whose site is for some professionals to exhibit and share their results and ideas, cloud computing is a matter of course. One of their sites is called action method, which allows users to upload and share files, including documents, audio files, and movies.

"When we started creating action method, we didn't know how many users there were, so we needed to make sure there was enough storage space to accommodate any number of users." "said Chris Henry, CTO of the Behance company.

Behence's website is owned by Rackspace. Rackspace's mosso Cloud files provide unlimited storage space and cost far less than the cost of San or NAS storage. Henry said: "With cloud storage, there will never be any shortage of storage space." ”

At the time, the hardest part was writing an API to access cloud files, because the Rackspace rest (representational state transmits) API was not available at the time, and Behance was too advanced. ' If we were to wait, it would be easy, ' said Henry, because the Rackspace Web Services API was pretty good.

In addition, Henry said, "Any cloud services to be integrated with an application, will require a certain amount of work." He advises peers to integrate cloud storage into their applications as early as possible so they can take advantage of the system without the storage constraints.

Over the past three years, Amazon has made great strides in its services, based on the company's own architecture. The company, headquartered in Seattle, said S3 had only 800,000 storage targets at the end of the second quarter of 2006, rising to 5 billion in a year, to 40 billion by the end of 2008.

S3 has a broad customer base, which includes individuals who store private data, large businesses that want to back up databases or archive data. For example, the New York Times uses Amazon S3 to store and send archived articles; Nasdaq uses Amazon S3 to store market history data, and businesses can access the data through the harsh replays tool.

The most tempting place for Amazon S3 is its cost price. Amazon S3 is priced at 15 cents per GB per month, and if customers use up to 50TB or more of storage, there will be price concessions.

"We're positioning Amazon S3 as an internet store," said the general manager of Amazon S3, "it can satisfy any HTTP request anywhere." How much to use, how much to pay. No capping, no limits. ”

Amazon S3 also needs a Web startups (network startup company) or a business customer to do some development work. Henry said the company had no contract for development but provided a list of partners. These partners have built their own applications on top of S3, S3 can support rest and soap interfaces.

Nirvanix is also a cloud storage vendor, and they use different methods. Their software cloudnas can provide 2TB or more space for users free of charge, without the need to develop APIs.

Cloudnas users can connect to a virtual NAS gateway from any Linux or Windows Server and then lead to Nirvanix SDN (Storage ibuyspy receptacle). The software allows access to SDN storage via standard storage protocols, such as NFS, CIFS, and FTP.

Nirvanix also has a unique feature of its Internet media file system, which allows users to manage their accounts. Nirvanix focuses on businesses that need to store confusing data, such as audio, video files, medical images, or engineering. "This information is not suitable for storage in traditional San or NAS systems," said Nirvanix CEO Jim Zierick. ”

Latency is one of the fatal vulnerabilities of internet-based storage. To this end, Nirvanix's SDN set up storage points in several places to reduce delays, including California, Texas, New Jersey, Germany and Japan.

Rackspace's Mosso is further integrated with CDN (Content IBuySpy Receptacle) through collaboration with Limelight NX Company. Users can get capacity from the nearest terminals on the Cdn.

"We already have a large number of users who use cloud files primarily to make full use of CDN." "says Johnson, chief architect of Cloud files. He also cautioned that Amazon's Cdn service, called CloudFront, does not necessarily have to be purchased with S3. "If you consider cloud storage, I think you will choose CDN." ”

Future prospects for cloud storage

According to storage Switzerland's Crump forecast, the situation in the next two years will be more beneficial to cloud storage vendors. However, there may be some challenges for customers. They need to sift through the flood-like cloud-storage new products in the market.

"In the area of cloud storage, the entry of some traditional it vendors has become a major event this year, such as IBM, HP, and possibly Microsoft." They will be more specific about their products and how to integrate with other products. Villars, of IDC, wrote in Ane-mail.

There are also reports of hype about Google's possible entry, but it will not be the main vendor of cloud storage.

Vaultscape Company has released a dedicated cloud for enterprise archive storage, mainly for companies with larger data volumes. Vaultscape has an experienced team with the founder and CTO Chris Williams (who served in Nirvanix) and technical advisor Aloke Guha (Copan BAE Company and developed its own file system and architecture).

Vaultscape's differentiator includes a self-healing file system and a "multiple-master Architecture (multi-master architecture)", with two independent vault available at all times. According to Michael Witz, co-founder and chairman of Vaultscape Corporation, the data will be replicated to two different physical addresses, but only through a core, so there will be no point of failure.

"Vaultscape shows us the growth of the industry, and the storage cloud is designed to meet the specific needs of the storage environment. "A specification is unlikely to suit all requirements," Witz said. ”

According to storage Switzerland's Crump forecast, the situation in the next two years will be more beneficial to cloud storage vendors. However, there may be some challenges for customers. They need to sift through the flood-like cloud-storage new products in the market.

"Cloud storage will be one of the technologies that benefit from the recession," Crump said. "For companies with budget constraints, the need to buy storage capacity is clearly more convenient and efficient than a one-time purchase of the entire storage system." ”

The real killer of EMC in cloud storage is architecture, what they call "cloud optimized storage (cloud-optimized storage)." EMC's new product Atmos allows enterprise IT departments and large Internet or telecommunications service providers to build their own cloud architectures.

Product management Director of EMC Cloud Architecture group Jon Martin says cloud-optimized storage is characterized by large capacity (petabytes, even higher), policy-based management, and operational efficiencies.

"The fundamental idea we're looking for is that storage technology is no longer confined to a single data center," Martin says, "based on the Atmos architecture that will spread across the globe." ”

In addition, companies in the cloud architecture sector include Bycast, Cleversafe, Ibrix and Parascale.

Such products are attractive to businesses that want to take advantage of cloud technology and worry about security. Because they can use these products to build their own cloud storage architecture inside the firewall.

According to Terri McClure, an analyst at Enterprise Strategy Group, the private cloud will grow as the web evolves-from a small token network within an enterprise to a virtual private network (VPN) to an IP network. "People may need to adapt to cloud storage for a long time and adapt to throw everything they have into the public cloud." "she said.

However, some analysts believe that cloud storage will grow in the next few years.

Edit Recommendation

What is the difference between cloud storage and cloud computing? What should be taken into consideration when deploying cloud storage?

With some analysis and user feedback, we can believe that hosting online cloud storage using data storage is more secure than a data center using a private storage platform.

Rebecca Wettemann,nucleus Investigation Co., Ltd.

• A thorough assessment of how difficult it is to retrieve your data before you log in and register

• See if there is a dedicated tool software for migrating data

• Determine if a separate charge is required to clear the data migration

• See if the special terms of the contract include data migration

Henry baltazar,451 Group

• Ask for an additional clause to be sent back to you at the end of the contract with your data on a removable hard drive.

Ben WOO,IDC

• Try not to fully use outsourced storage. Using a hybrid strategy allows primary storage in your internal organizational architecture to react faster

It's like asking yourself. Cloud Storage Services Data Center: How do I make a daily backup? How do I save backup files?

Greg Mcgovern,adventist Tiyatien Chief Technical officer

• Ensure that the services they provide conform to your data management strategy

• Use open file format if available

• Daily practice to ensure you can retrieve your data

• Ensure that you have access to the appropriate service level identification to meet the needs of your data application

• Never outsource unless you can be sure that others can do better than you.

What is the difference between cloud storage and cloud computing?

Cloud computing is the development of distributed processing (distributed Computing), parallel processing (Parallel Computing) and Grid computing (grids Computing), which automatically splits large computational handlers into countless smaller subroutines through the network, A large system of multiple servers is calculated and analyzed to pass the processing results back to the user. With cloud computing technology, Web service providers can handle tens of millions of billions of dollars of information in a matter of seconds to achieve the same powerful network services as supercomputers. The following figure:

　　

The goal of the cloud computing system is to run on the PC, or stand-alone, personalized operations on a single server migrate to a large number of server "Clouds", which is responsible for processing the user's request and outputting the results, which is a system centered on data operation and processing.

Cloud storage is a new concept that extends and develops in the concept of cloud computing (cloud computing), which means that a large number of different types of storage devices in the network can be assembled together through application software through the functions of cluster application, grid technology or Distributed File system, A system that provides data storage and business access functions together externally. When the core of cloud computing system operation and processing is large amount of data storage and management, cloud computing system needs to configure a large number of storage devices, then cloud computing system is transformed into a cloud storage system, so cloud storage is a data storage and management as the core cloud computing system.

Compared to cloud computing systems, cloud storage can be considered a cloud computing system configured with large storage space. The following figure is a comparison of the architectural models of cloud computing and cloud storage:

　　

From the architecture model, the cloud storage System system has one more storage layer than the cloud computing system, and there are many functions related to data management and data security in the basic management, which are identical in both the access layer and the application interface layer.