In the big data age, the development of PC Internet, mobile internet and social platform makes the data appear geometrically and explode. With the increase of data source channels, the data of the network is no longer a single data type, a large number of unstructured data has changed the composition of the data model. Now big data makes the entire Internet technology faces a new round of shuffle, how to use large data will become the key to determine the future fate of the enterprise!
So how big is the big data? Many people may ask this question. In the 2013, data stored in the world are expected to reach about 1.2 ze (about 1.2 billion TB) bytes. What is the concept of 1.2 billion TB? If all this data is stored on a standard disc, these discs can be stacked into five piles, each of which can reach the moon. And if this data is stored in 1TB capacity of the hard drive, then will need 1.2 billion pieces, these hard disks can be divided into five piles, each equivalent to the length of the Great Wall (6700 km).
Cloud storage is an emerging storage model that accompanies cloud computing and is a relatively unfamiliar area for many people. In fact, cloud storage technology has long been deep into our lives and work, but many people do not realize its existence, the current cloud storage technology in large data solutions have some cases. With the advent of the big data age, cloud storage will assume more tasks and missions
Microsoft SkyDrive Cloud Storage
Recently, the introduction of the 360 network and Baidu 1TB competition to upgrade the incident I believe we have heard! Yes, these disks are really the most typical cloud storage applications. These cloud storage service providers to provide users with quantitative storage space, so that users can achieve through the PC, mobile phones and other terminals to access demand. Network disk, cloud disk These cloud storage technology is the most typical and the most elementary cloud storage technology, now some enterprises have the cloud storage technology has a deeper development and innovation, such as the thunder of the cloud acceleration function.
2013 Internet Conference Thunder CEO Shenglong The Cloud storage acceleration brought about by the change described as: "Let big data like water flowing freely." "At that time when I heard this sentence, I do not understand this sentence, as far as I know, the Thunder Walkman this typical cloud storage model has not been successful, and is about to stop support on September 1, 2013, then Mr. Shenglong in the General Assembly repeatedly stressed the Thunder Cloud acceleration technology? And how can large data flow freely?
Thunder has a deeper application to cloud storage technology
With these puzzled, I looked at the Thunder cloud accelerated some of the business. At present, the Thunderbolt already has high-speed channel, offline download, Thunder cloud broadcast, Thunder online games Accelerator, mobile phone thunderbolt and many other products, and covers the PC, mobile phones, flat, TV four screen, can say that the Thunderbolt now basically completed the overall layout of the business.
According to the author's research on these services, many services, such as off-line downloading, high-speed channel and so on, are based on cloud storage, which can be said to be a kind of innovation and evolution of cloud storage technology. Thunder to cloud storage technology has a deeper application, not just limited to the traditional data access and data protection, but also the application of cloud storage technology to download, video, netizens and other fields, is the cloud storage technology application to large data on a new innovation!
Many kinds of thunder business
Thunderbolt using cloud storage technology to ensure that users can obtain a smooth, coherent experience, but also launched the cloud based on the account to achieve the function of synchronization, it can be said that the Thunder has now been using cloud storage to achieve the entire product layout, so that large data transmission between the more free. So how does thunder use cloud storage technology innovation to promote the flow between large data?
Below we download the business through the offline to see the explanation under the Thunder Cloud storage innovation place.
Offline download is an important function of thunder cloud acceleration, when the download time to encounter the remote resources, due to the complexity of China's network resources, if downloaded from the original server to the local, may be due to the lack of resources, distance and other reasons so that the download speed is not high, so the use of ordinary download will waste a lot of time and electricity. And when the resources are good, high speed, it will affect the user's normal network problems, resulting in other tasks can not be completed.
Thunder offline download every transmission of a file, will record the address of the file resources, CID value, as well as other characteristics of the file information. So accumulate, the Thunderbolt will save a large number of file characteristics information, when another user downloads the same file, the Thunder Cloud Accelerator service no longer needs to download files from the server where the resource resides, but is copied directly from the resources within the Thunder Cloud Accelerator to the user's offline transport folder, a few gigabytes or even dozens of GB The large file in an instant on the transmission completed, so many people think offline download extremely fast, called ' seconds Kill '.
What kind of support does the cloud storage background need?
"Second Kill" download is not Thunder server powerful, but the Thunderbolt use cloud storage technology of an innovation. This saves a lot of bandwidth for transmitting the Thunder cloud to accelerate new resources that do not exist on the network. It is the use of such methods, the Thunderbolt constantly accumulate resources, constantly indexing resources. But for a long time no one has downloaded the file, the Thunder Cloud Accelerator service will also delete this file index information, reduce server space occupancy.
Offline download services enable fast transmission of data between people, and in the age of large data, make the data flow more cheerful. But the current Thunderbolt offline download only for members to use. If you are a member of the user, then you can rely on the thunder of the server's powerful download function download files, or directly from the Thunder server directly download someone else has downloaded the file, is equal to copy and paste, which saves a lot of time, promote the flow between the data.
Cloud storage servers require strong background support
The current thunderbolt per month as high as more than 4 million members, and in more than 4 million of users behind there are nearly 300 million of free users, these users can also use the offline download business thunder, which must require a strong cloud storage platform support, to provide users with off-line download business, It also meets the challenge of increasing the volume of data files and the proliferation of user traffic. At the same time, cloud storage platform needs to continuously storage capacity of its server platform, System broadband, I/O needs increased.
Below we will discuss what kind of platform to meet the needs of the Thunderbolt.
To build a cloud storage platform that meets the needs of users, we first need to understand the principles of an offline download platform:
(1) The user submits a download request through the client or the Web interface to the server.
(2) The server side accepts the request, first inquires whether the user submitted download link has been downloaded. If so, put the downloaded data file (or just the link to the file) into the user's server-side online space. If not, turn on multithreading to download (or use a company-specific peer-to-peer approach).
(3) After downloading, the user logs in online space and retrieves the downloaded files. In the meantime, we can use the Peer-to-peer method provided by Thunderbolt to obtain data from users who have downloaded or are downloading the same files.
(4) Offline download more for the unpopular resources, or less resources of the document. When the server side is not downloaded for the user, the user also needs to download the files from the server using the downloading software. Compared with the direct download, it increases the speed of downloading resources and saves time.
How to build an offline download support platform?
So how many servers to meet the Thunderbolt 4 million members and nearly 300 million of the average user needs? The author began to mistakenly believe that this business will require a lot of server support. But after the thunder of the offline download platform for research, found that support offline download server is not as much as we think. After all, these users can not download the data at the same time, and each user's download is not all saturated, unpopular data is not always stored, so the server needs are not as much as imagined.
Let's take a look at the composition of this platform, and perhaps you will understand why it is not necessary for the server to meet the needs of users.
Offline platform needs to download server, Web server, data storage server three platform to support offline. These three platforms require excellent performance support, and there is no difference in processor and memory, and the main difference is in the storage drive.
Here we can predict the storage requirements of three systems from the price/performance and requirements:
Offline download server mainly provides speed download, so the server's hard disk performance and capacity requirements are higher. Select SAS hard disk, the data storage server provides storage capabilities, storage capacity requirements, but not too much demand for speed. The use of SATA hard disk, the Web server to the speed and performance requirements of the hard disk, but the size of the capacity is not too much requirements. SAS hard Drive is satisfied.
With the improvement and innovation of cloud storage technology, Thunderbolt solves the problem of support of off-line downloading business. At the same time, this technology can be competent for a variety of business, such as data storage, database applications, high-performance computing, computer clusters and other fields. Cloud storage technology makes data storage simpler and more flexible. In the advent of large data, cloud storage technology innovation and development will lead to a new revolution in large data, so that the real realization of the data flow freely.