Briefly illustrate cloud computing with examples of life

Cloud computing began to be widely recognized by us. A large number of terms such as SaaS, PaaS, and IaaS began to emerge. Many people may say that these are good ways to save money by using IT services. Indeed, the rapid growth of server virtualization and Internet technologies has led to a gradual increase in the use of IT resources. Of course, we can use the same approach as "carpooling" and "group buying" to maximize the benefits. Cloud services on the "buy" and "carpool" can be better than in real life version done. First of all, according to the amount of resources used by each customer and the use of time for different charges. You can also make a corresponding dynamic adjustment of the size of "guest room".

Since the main task of cloud computing is off-premises execution, the required capacity for computing is not limited by the amount of computing resources available in a given place. As a result, scalability becomes a natural feature of cloud computing. This flexible scalable feature is very useful. Last year, the online Olympic ticketing website crashed during the opening hours of the service. That unfortunate incident happened precisely because the hardware and software architecture of the server that supported the ticketing website in the back-end data center at that time did not work out a flexible and scalable way of working. When the number of service requests far exceeded the server can provide the workload, the ticket site had to "strike."

Another very important feature of cloud computing as a service is the provision of value added services to users. What is the value-added services on the cloud? Let's imagine an example of such a cloud storage. If there is such a service provider, it has a very efficient data compression algorithm and deduplication algorithm. As long as its algorithm is effective enough, the service provider can provide scale cloud storage services to end users, but actually do not need to equip their own storage devices with the corresponding scale. Because it can do its own handling of user data itself - fully compressed or deduplicated - and then subcontract the processed results to other cloud storage service providers such as Amazon S3. So this clever cloud storage service is itself a user to rent other cloud storage service providers. Its advantage over the latter is that it takes away the cost of having hardware storage resources apart from the size of the packet forwarding packet. Together, these two components form a distinct advantage for this clever cloud storage provider relative to other third-party cloud storage providers, making it a profitable space for storage services. This clever cloud storage service provider is actually doing a value-added storage service. From this example we can also see why cloud computing / storage is often referred to as third-party services.

It is generally accepted by the industry that enterprise private clouds are widely used before public cloud. What is the enterprise private cloud? Roughly on the internal IT resources to do some reorganization and rationalization of the results. The historically developed IT systems within the enterprise not only have huge waste of resources (for example, each employee is equipped with a set of both heavy-end front-end hardware and expensive software license, the vast majority of applications only to deal with office Desktop office work), but also on each front-end machine must operate on-premises (on-premises execution) management and maintenance, in order to catch up with increasing manpower costs. On the issue of cloud security, the implementation of enterprise private cloud is simpler and easier than that of public cloud, so the private cloud of enterprise takes the lead in public cloud.

Enterprise private cloud is not a real cloud computing? Personally, I personally think that if a privately owned cloud is used on a small but well-established basis, its private cloud is equipped with its own storage disaster recovery and redundant backup system (according to the Network Storage Industry's SNIA standard, redundant backups must be deployed in at least three different geographies Location), then such a private cloud should not be seen as a strictly cloud computing, because it is not in place for the rationalization and reorganization of IT resources, incomplete, still with local heavy configuration, high cost, lack of flexibility and the like Typical non-cloud computing nature.

A more rational and thorough private cloud approach should be to have multiple private clouds share a common cloud storage service. This shared, common cloud storage service places its resources among many private clouds and can therefore be called a "private cloud". Not only does it provide dedicated storage services with redundant backup for enterprise private clouds, it also provides cloud computing users with a convenient "move" service to prevent so-called vendor lock- in phenomenon occur. In terms of hardware architecture, such a "private cloud" is more than enough for iSCSI-attached disk arrays because cloud clouds do not require real-time high-speed data processing and do not require very high-end storage technologies such as Fiber Channel attached disk arrays. However, in the software architecture "private cloud" requires a highly intelligent, policy-based SaaS service level and information management capabilities, in particular the need to achieve strong security data isolation protection technology. From a purely technician point of view, I do not think there is a cloud storage solution on the market that offers the cloud storage services of the quality required for "private cloud storage clouds," mainly because of the poor security of the software architecture A success. EMC's Atmos onLine should have the earliest possible upgrade to software architecture.