Virtual cloud practice sharing within large enterprises

Chichu, senior project manager at Microsoft China Cloud Computing Innovation Center, shared the "virtual cloud practices within large enterprises".

Why do companies do cloud computing?—— save money and make money

Chichu a question to start his own speech, why do companies do cloud computing? For businesses, he says, it is about making money and saving money. The private cloud can basically save money, can be a lot of physical machines into virtual machines, the virtual machine combined, in the virtual machine dosage is not so large, the use of virtual institutions, and then when the user back to the virtual machine open, so you can save money, machine reduction, IT staff can be reduced. The part of making money is the common cloud. Put the application to fair, users can subscribe to use. For example, office365 is a subscription to the use of services, users need to use the time to buy this application, put it on your terminal.

In the next two cases, Chichu, through the sharing of actual cases and his own experience, elaborated on a number of new ideas: How to let the private cloud from saving money, into making money, aroused the interest of participants.

Description of requirements and ideas:

A clothing manufacturer in the major stores and large shopping malls will have retail outlets, each retail point is very simple, there are shop-length or more senior managers can SS. But if the computer has a problem, to send people to repair, not only a long time, and the need for the human cost is very large, so they asked to build a cloud environment.

Cloud computing means that all calculations are done not locally, but on the cloud, and when each branch is using the app, it does all the computing work in the cloud. Of course, VDI is a very straightforward approach, but if there is a network interruption, VDI can not go to access the system, so Microsoft will have other technical--hyper-v, in the local first calculate, wait until the network unobstructed, and then upload the data to the cloud.

The concept is very simple, but it is not difficult to do. However, because different products are not the same factory to do, there are responsible for proofing, materials, design drawings and other work partners.

▲ Retail Industry Cloud Platform Application scene map

As shown in the figure above, because the partner has done the design, after the company to decide which to purchase the design, so you have to do some of the system Information integration, partner development environment to unify. In this way, proofing after playing well, download to the machine will not need to set the machine on the fine step of the place, just put the design map in, you can know how to cut the cloth, lines to sew a few lines, so the development environment must be consistent.

Also, when there is a new product line, and not every partner to change its application, so from the partner point of view, also want to put applications in the cloud. When there is a new partner to provide the development environment to it, just povision a few virtual machines to it, the application in the virtual machine has been installed, and then just have to tell it how to connect in, how to work in the offline way.

Similar to this mode of operation-"manufacturing (manufactory) + retail" is to do their own to do more than one company, many companies are such. So, it also wants to put my system in place, the IT staff learned that after the platform was developed, they could set up a consulting firm to sell it, and in fact, it did set up a consulting firm to sell the technology to someone else who did the same thing, so from a commercial standpoint, It changed from private Cloud to public Cloud.

Solution Architecture and Technology implementation:

▲ Retail Industry Cloud Platform Architecture Chart

As shown above, from the underlying hardware, middle-tier products, and then management. Chichu Rong believes that the most important thing is to follow the customer's own original management process, their original it habits tool, try to do not let it to change its tools, not to adopt the cloud, the original IT management model all to get rid of it, this is absolutely not recommended practice.

Description of requirements and ideas:

The second case of a large enterprise is a commercial software company in Beijing. They want to do is very clear, is to sell software, but there is no way to predict at a specific point of time, whether the number of 10,000 simultaneous users, or 10 million, it wants to have a flexibility to do this thing, and then very clearly tell us that Don't want to be able to estimate the amount of traffic because there is no way to build a large data center, to meet a now do not know when it will happen.

Based on this demand, since only focus on the application, the final solution is to put the application on the Microsoft's cloud platform--azure, the biggest advantage is that the cloud platform is more resilient, and can save a lot of data center maintenance and construction costs. Of course, we know that Windows Azure is not yet on the mainland, but it is revealed that Azure has a hope of landing in mainland China in 2012.

In addition to the above requirements, they have a "data market" idea. Commercial software, database is the most important, contracts, customer information, etc. are the most important. From the perspective of this business, all registered should have a centralized database to manage, so I was doing a lot of data search, you can put all the user data are over, but now a big problem is, this massive storage to find fast enough?

For example, if you buy a car, register your name, phone, etc at a car dealership, the dealer will put the data on their own server, which is actually just a client's access information. But if a company is looking for people aged 40 to 45 who have bought a car in China to live in Beijing, if the search is only a distributor, then this data is not representative of the meaning. It wants to be able to find every different vendor, so it's necessary to coordinate the vendors to share their data.

So how to reconcile these data and sell it to people who need it, which is called the data market. They want to put all these things on the Azure app, and not over the Internet, so it's slow, but hopefully everyone will share that part of their data in a centralized database environment.

▲ Cloud Platform architecture diagram for commercial software vendors