Four key to cloud services

Cloud computing, the fourth wave in the IT industry, is coming at a rapid pace, with lower costs, more accurate cost-matching and no overinvestment problems making the cloud more attractive.

In addition, competition between cloud service providers if the market is developing well, it can become quite intense, and this competition also makes cloud services a price advantage.

However, for businesses, whether public or private, the key remains the ability to manage effectively. For the cloud model, there are four main factors in the internal configuration that can help you choose the right strategy to monitor cloud usage efficiency, dependencies, and profitability. Here are the four factors.

Network situation

"For the cloud model, there are four main factors in the internal configuration that can help you choose the right strategy to monitor your cloud usage efficiency." ”

For cloud systems, the network is used in a different way than traditional computing. For example, the way a midsize enterprise uses its own services to use a network is completely independent, meaning that departments and employees have access to audited, specific external Web sites and data sources. For companies that do the same work in cloud mode, they need to pay more attention to the traffic problems, which is very similar to the traditional point-to-point leased line.

Since not all Internet service providers can respond effectively to such traffic, it is important for enterprises to monitor parameters such as network latency, throughput, and reliability. The reconfiguration of the network configuration can have a significant impact on connectivity performance, and the processing of data backups can be more expensive than the collection.

Storage requirements

As with all other parameters, the key to understanding the storage requirements in the company's actual work is to master the core of usage patterns. Storage is a very sensitive part of the cloud environment, and demand is growing over time, and, from a vendor's point of view, prefers to recommend more flexible options for more profit.

Of course, if the company chooses to build a private cloud, it needs to address all types of standard issues in the self-configuring storage, as well as the need to meet flexibility and scalability.

Here are a few key parameters to focus on:

Ø availability, should be clearly defined in the service level agreement (SLA) and need to include a comprehensive assessment approach

Ø reliability, which also needs to be planned based on the actual situation

Ø you should adhere to the principle of simple application of API connection settings that can be used on a web-based service

Similarly, other aspects to be noted include:

Ø How the connection is provided (how the user should respond to the performance problems that arise when the filing system is compromised)

Ø Billing Mode (according to what standard charges, how to confirm the use of time and location, how to limit)

In general, the most needed attention in the system remains the part responsible for security.

For the security model in the former cloud environment, there is no common standard in the industry at present. As a result, no one will think that the reality will change dramatically overnight. In fact, the original standards can still be used, there is no understanding of the time between the enterprise and the cloud provider to communicate security needs, for example, how to ensure system security, how to check applications, and what lines should be used in a state of emergency.

Processor Standard

There are many different options available for vendors of infrastructure, the service (IaaS) class. The reason for this is that there are currently no viable industry standards to measure computing power.

For the security model in the former cloud environment, there is no common standard in the industry at present. As a result, no one would think that the situation would change dramatically overnight.

As one of the oldest vendors, Amazon EC2 Service uses the standard called the Flexible Computing Unit (ECU). It is equivalent to a 2006 standard vintage 1.7GHz Xeon processor in terms of performance. The standard used by Wei Rui is called Vpu, which is about the equivalent of a 2.9GHz microprocessor without accurate testing. Other vendors use models based on the number of processors or GHz to sell.

Since these standards are based on virtual processors, the reality is not static. After a user chooses to purchase a certain amount of processing power, the cloud service provider's management program assigns the remainder to use by other applications. Such an allocation may result in actual work being handled, but the application of internal activity shows that everything is working properly.

At this time, the work we do is to test the idle time of the processing capacity, if it reaches 20%, it indicates that more resources need to be invested. Alternatively, we can also choose another increasingly popular measurement standard, CPU embezzlement rate. This percentage indicates how much of the "own" CPU is used elsewhere.

Of course, there is no substitute for performance. As with physical servers, the choice of specific criteria depends on the objective situation in the actual work. Since there is no real cross-platform standard at the moment, in order to ensure that the parameter does not cause a problem, the SLA needs to be regulated and the monitoring measures need to be included.

Energy model

As the last parameter mentioned, it belongs to very unique. This is a major constraint for data centers, single CPUs, and single server energy consumption, and is a key concern for cloud service providers. In addition to the overall cost, it is the least part of the vendor's willingness to communicate with customers.

From the point of view of energy utilization, it is entering into the era of intelligent mode leading. Companies will need to take responsibility for reducing energy consumption and hope to gain a first-priority advantage from closer attention to technology energy. For the cloud service provider, this means that the customer should be provided with the appropriate information to help the enterprise to adjust the internal calculation. But the reality is that they will not do so unless the client makes a request.

"Edit Recommendation"

Cloud computing and who will be the fashion to choose public cloud services before the analysis of its advantages and disadvantages cloud computing the next seven popular IT jobs why is the dilemma of cloud service choice? The secret of cloud computing success: Understanding the SOA architecture correctly who can master the SME cloud computing tomorrow? "Responsible editor: Xin Wei TEL: (010) 68476606"