Strictly speaking, cloud computing is not a new term. It can be said that it is neither a new technology, nor a new concept, even if the word cloud computing itself is appropriate is also a problem. The concept of cloud is more than computing. This is an overall guideline and future trend in the field of IT system construction. This represents a brand new IT era.
As early as the beginning of the 21st century, IBM had proposed that On Demand be changed as needed. HP also proposed Utility Computing's utility calculation, and H3C proposed IToIP. Even in the mid-1990s, there have been a number of pioneers operating around the world with ASPs (application service providers) and SSPs (storage service providers) as pioneers in cloud computing And practitioners. The above concepts and business ideas do not differ much from those in today's cloud computing, except for the different perspectives on the same IT development perspective.
So what is the vision? Just hope that IT resources will one day be as "plug and play" as the electricity and tap water used today. These IT resources include applications, software, hardware facilities, and more. We have never heard of "cloud power" or "cloud tap water." Because there is never any need to be concerned with generating electricity, power-supply equipment can easily be accessed using the technology and management of power transmission equipment; clean tap water resources can be used without the need to care about the extraction, purification and delivery of water. Cloud represents the general trend of socialization, service and specialization of resources.
The question we are going to discuss is: Are there any IT areas that have been resourced or "cloudized"? What has prevented IT "cloud" implementations? What do we need to embrace the "cloud"?
If we talk about IT in general and voice communications in IT, we have already implemented "cloud phones" in both fixed and mobile phones. We only need to apply for services to operators according to consumer demand, you can buy the terminal as you like to access the use of communication services. If we go back to 20-30 years ago, we can also see that every government agency or enterprise is buying a PBX or even setting up an operator department for communication. The emergence of large-scale digital program-controlled switches and telecom operators has enabled this kind of self-cultivating communication construction to enter the era of "cloud communication" of social collaboration.
We are familiar with the Internet is more inherent in the spirit of globalization division of labor collaboration, from the outset is the attitude of network resources and services to create a "cloud network." Even though some of our industry sectors are still bored to build an independent vertical network for internal use, at least nobody ever wants to build a "own" Internet.
Cloud is never new in the IT world, and we use it every day in the fastest-growing areas of voice and data communications. From the history of human development, deepening socialization and specialization are the major trends in social development. In the IT industry, this is no exception. Therefore, we should not look at this floating "cloud" from the perspective of "new technologies" or "new concepts." Instead, we should regard it as the general trend of IT system construction.
Now to answer the second question, what are the factors that hinder the "cloud" implementation? There are two main reasons:
First, the traditional management system of information resources to occupy, artificially hinder the flow of information and sharing. Without the "water" of the mobile cycle, "cloud" naturally can not be generated.
Second, the rule of law in the IT development process and the interests of the leading, leading to many standard. Many vendors are defining proprietary application standards and data formats to lure or force users to build their IT business in the context of IT system building. Eventually lead to business fragmentation, system chaos.
A simple assumption is that if a user's machine room runs 4 to 5 servers, operating systems, databases, storage, network devices, and security devices from different vendors at the same time, the application has 3 to 4 different developers Their own fight, never in accordance with the norms of global information standards for information definition and system development, the data center is completed only to physically pound the data into a room, the business is still fragmented, irrelevant, Then, as the user faces more and more severe business continuity, security level protection, cross-business integration, emergency command and so on new business theme test, it is hard to avoid getting into exhausted state of fire and the business is naturally difficult A higher level
Therefore, the cloud has important practical significance and value for us to reflect on the IT construction process and rethink the future IT road from the perspective of systematic, intensive, integrated and social professional division of labor. We do not need "cloud computing" hat, but we can examine our building ideas accordingly: What is in line with the development trend of the practice? No longer under the banner of information technology, the message is locked in the box at the bottom of things.
For management issues, we should realize that: one of the premise and result of full informationization is the flat management, which is the customer-oriented or service-oriented management. We are delighted to see that following the remarkable results achieved by the integration of operators' BOSS system and the consolidation of large-scale financial data businesses, many government departments are also striving to integrate management process changes and IT systems. From the public security department of the police comprehensive system to the school's digital campus; from the social credit information system, one-stop government services to the site integration. Instead of focusing on the realization of individual functions, more and more government and enterprise information systems focus on the big business of large systems, break down the barriers to division of labor within the internal departments and even break the "digital divide" among different industries or enterprises and are actively To the standardization of information, information exchange and sharing, business process integration direction.
On the basis of management optimization and business integration, the reconstruction and integration of technology and system become the second problem to be faced. In front of the analysis, hardware and software technical standards, data format standards disputes and interests of manufacturers dispute is caused by today's system integration and the root cause of the difficulties. From the current stage into the "cloud era" of similar communications and networks, there are at least three more technically significant steps to go from bottom to top: full standardization of computing and storage technologies in hardware and management technologies; operations System fully standardized, with a high degree of interoperability and on-demand deployment capabilities; The main application of the data format is fully standardized and highly interoperable capabilities.
The process of spanning the three steps requires a large amount of middleware. Virtualization technology breaks the hardware and software structure that was tightly bound together and becomes a resource-based IT architecture that provides service layer by layer. This is often done in cloud computing Mentioned PaaS (Platform as a Service) / SaaS (Storage or Software as a Service) / IaaS (Infrastructure as a Sevice) value.
Imagine such a scene: a business office, go to work at 9 o'clock in the morning, the staff is busy receiving and dispatching mail, at this moment, can gather 80 among 100 computational nodes to concentrate on the mail business; After 10 o'clock, as the terminal customer enters the business gradually Places, business systems began to peak, the mail business began to decline. Administrators and even systems can automatically release up to 70 quickly available support for business systems to the 80 compute nodes used by the mail system. All computing resources like a pool that can be dynamically reconfigured to provide timely support and flexible services to tail-end fish for different applications.
Imagine a scenario in which about 10PB of data (10PB = 10,000TB) is managed in one data center, with 100TB of real-time online transaction data and 1,000TB of data warehouse and decision support system, including images, videos and Approximately 2,000TB of common text file data and nearly 7,000TB of other historical data are stored for 30 years.
Due to the full realization of the storage of standardization, intelligent, virtualization, service, data resources management very easy. Each server simply does not need to pre-allocate space, and only needs to send data to the storage service platform to read and write data. The storage platform intelligently determines data according to different data read-write requirements put forward by the front-end thousands of computing nodes Service Strategy:
1, with high performance requirements online transaction data, the storage system automatically deposited into the capacity pool composed of solid state drives to provide high performance data read and write services and high reliability RAID10 protection services, while the capacity pool has been automatically set up a backup, the same city , Off-site multi-level disaster recovery data protection capabilities; also allocated enough IOPS capacity to serve the data read and write operations.
2, for data warehousing and decision support systems, the storage system automatically allocates data to the SAS hard disk configured with higher performance RAID5 for lower-cost configuration and data storage, only one of the important metadata for high-level backup And disaster recovery protection, process data only simple near-line backup protection.
3, for nearly 2000TB various unstructured data, according to the data itself describes the data format, timeliness requirements, transmission bandwidth requirements, the storage system automatically divided into different data areas assigned to different media, providing different bandwidth conditions, different data Backup and disaster recovery capabilities respond to different data access needs.
4. According to the business rules, the storage system automatically archives data with access frequency of less than 10 times in one year or more to the green data archive area with MAID (Massively Inactive Disk Array) technology support. The hard disk is normally powered off, Save on archiving energy, automatically power-up related media for service only when there is access or new writes. If the data is not accessed for more than a year, the storage system will also automatically archive this portion of data to tape for offline storage.
5. At any time, storage systems automatically combine and allocate storage resources based on access characteristics and data characteristics, and continuously provide administrator with usage characteristics and record analysis of various data areas. Once a data area reaches the threshold of capacity, performance, or safety, only the standardized media, performance processing units, etc. need to be standardized according to the standardization rules, and the storage system can automatically add new resources to the resource pool.