Shulou(Shulou.com)12/24 Report--

Today, our country has entered the era of computing power. In the era of computing power, computing power is the center and the network is the foundation. The network runs through the whole process of production, transmission and consumption of computing power, which requires efficient release of computing power, flexible scheduling of computing power, and optimal service at the edge, so that thousands of industries can use computing power as easily as electricity.

At the same time, according to IDC, more than half of the global economy will be based on digitization in 2022, while some research institutions predict that more than 95% of enterprises will work entirely on cloud-based platforms by 2025.

Especially this year, with the popularity of the AI model represented by ChatGPT, the demand for computing power is also showing a geometric multiple growth. Wang Jian, academician of the Chinese Academy of Engineering and founder of Aliyun, said at the 2023 Cloud Conference: "the relationship between cloud computing and GPT is the relationship between electricity and motor. In fact, the computing power of cloud computing will be consumed by these motors in the intelligent era, that is, models." The network naturally becomes the power grid that transmits electricity to the motor.

Convenience and flexibility have become the "must" of the network.

With the popularity of the mixed office priority work mode, videoconferencing has become an important office and production tool, accounting for more than 80% of video traffic on the Internet. In 2025, 90% of enterprises will deploy their business in the cloud, especially in manufacturing and retail. With the rapid increase in network size and complexity, 80% of enterprises have more than 20 operation and maintenance systems, and campus, wide area, DCN and other networks all rely on manual operation and maintenance.

In such a huge network demand, also put forward higher requirements for the network, before, the cloud network is more to solve the problem of connectivity, with the penetration of intelligence, the future network will develop towards a more intelligent direction. In this regard, Sun Chenghao, senior director of Ali Yunyun Network product line, said that since 2020, as more and more enterprises will put their applications on the cloud in a native way, the development of cloud networks has also entered the third stage, "Ali Yun believes that in the third stage, cloud networks mainly solve the cloud network connected by applications, facing enterprises to connect to more types of rich application needs." That's what Sun Chenghao said.

At this stage, enterprises need more convenient networking capabilities, as well as more flexible network architecture. Sun Chenghao told titanium media that from the perspective of the industry, most users are not cloud native. There are many basic users who have their own architecture and IDC computer room under the cloud. "when these users migrate to the cloud, they will consider the reliability of the network and the difficulty of building the network." Sun Chenghao pointed out, "Ali Yun believes that the cloud must be flexible and simple, and the export of this convenient networking ability through the ecological form will become the key to the next enterprise-level networking."

In Sun Chenghao's view, what enterprises need at this stage can be achieved through simple API operations to achieve minute-level network construction and configuration capabilities of the network. "convenience is one of the biggest needs for many enterprises." Sun Chenghao stressed.

In addition, with the explosive growth of the amount of data, the business demand of enterprises also presents the demand for high concurrency of massive data, which naturally puts forward higher requirements for the flexibility of the network and the ability to deal with high concurrency business.

With the continuous increase of cloud computing resources and the rapid development of business, cloud network needs to be able to expand smoothly to meet the needs of the future. Specifically, cloud networks need to support scale-out and scale-up to achieve seamless integration and efficient use of resources.

The strong resilience is also the biggest difference between the cloud network and the traditional network. Traditional networking needs to give priority to the purchase of high-quality routers, and at the beginning of networking, the specifications of routers determine the "upper limit" of network channels; in cloud networks, enterprises do not need to worry about the constraints of hardware equipment, but adopt a form similar to public cloud, which can be paid on demand to get strong bandwidth support in minutes, so that the network can really scale freely. " Just like we turn on the faucet to flow out of the water, the enterprise does not need to care about the thickness of the 'pipe', just choose the size of the tap according to the demand. " This is Sun Chenghao's analogy for cloud networks.

Not only that, cloud networks also need to be flexible to adapt to the dynamic changes and elastic needs of cloud computing resources. Specifically, cloud networks need to support virtualization technology, containerization technology, software-defined networks, etc., in order to achieve dynamic deployment and flexible expansion of resources.

Safety is essential.

Convenient deployment capabilities and flexible scalability are the two biggest needs of enterprises for cloud networks. Of course, on top of these two, there is also a focus that most enterprises in the industry pay attention to-security.

Data security has been a "clich é" topic, and it is also true in network construction.

Security is one of the core competencies of cloud networks. Cloud networks need to ensure the security and privacy of data and prevent unauthorized access and attacks.

For example, data in a cloud network may be at risk of being illegally accessed or compromised. Attackers can obtain sensitive data through network penetration, social engineering and other means, causing enterprises to face heavy losses.

For example, the current distributed denial of service (DDoS) attack is one of the common security threats when enterprises use cloud networks. The attacker floods the target server by controlling a large number of hosts, making it unable to provide services properly.

In the cloud network environment, through the virtual private network, provides each user with a logically isolated network, in this way to ensure the network security of users. However, with the development of cloud networks, in Sun Chenghao's view, simply providing a network that is isolated from each other can no longer meet the current needs of enterprises for cloud networks, "because enterprises will have the desire to access other networks through this network." in the cloud, every enterprise does not want to be in an isolated island after going to the cloud, and all want to better connect with the outside world. " Sun Chenghao pointed out.

In the process of connecting with each other, enterprises need a secure "boundary". "the overall security of the cloud network has two foundations. One is that there should be enough security in this isolated network; in addition, there should be access control and firewalls on the border to make the boundary secure, so that there will be no security risks when visiting each other." or minimize the occurrence of security risks. " That's what Sun Chenghao said.

Specifically, the cloud network needs to have data encryption, access control, security audit, intrusion detection, in-depth detection, external firewall and other functions to ensure the confidentiality, integrity and availability of data.

Wu Jiangxing, academician of the Chinese Academy of Engineering, also said at the 2023 Global Digital Business Conference that unlike in the past, the cyber threat is expanding from the level of information threat to the threat of "information + physics + data domain". There are a large number of unknown network attacks in all aspects of data collection, collection, transmission and transaction, and the economic losses caused by network security problems are very serious. therefore, the traditional underlying paradigm of digital ecosystem needs to be reformed urgently.

Of course, one manufacturer alone cannot provide comprehensive network security services. At this time, it is necessary to "warm up the newspaper regiment" and build an ecological circle of network security. It is understood that Aliyun has previously partnered with Fortinet to launch network security services that are available as a service and pay on demand. And this kind of strong combination to provide network security services is also very common in the industry. For example, China Information Communication Institute and Qingtenyun Security have reached a strategic cooperation to lay out the industry's cloud native security.

"Cloud network security essentially represents the integration of traditional security and cloud network itself, and such challenges and demands. Because these challenges and demands essentially come from many enterprises that rely on the security industry and manufacturers." This is Sun Chenghao's summary of the current situation of cloud network security.

Cloud Network that is constantly evolving

Under the basic premise of security, cloud network has been favored by various industries because of its convenience and flexibility in recent years. And the cloud network is constantly "evolving" with the tide of the times. With the development of intelligence, the intelligence of network has been put on the agenda. In this regard, Sun Chenghao said that intelligence is a tool, and in the field of cloud networks, the purpose of intelligence is to make cloud networks simpler, easier to use and more secure.

From the current situation of network intelligence, the realization of automatic scheduling through API is the embodiment of the more common intelligent cloud network. In response, Sun Chenghao told titanium media that the traditional call network requires people to participate through the console. "this efficiency is obviously more efficient than using program automation through the API port." Sun Chenghao pointed out.

Not only that, through intelligent technology, cloud network now has the ability to provide network "visualization" for enterprises, so that the network has the ability to be observed, so as to help enterprises better coordinate network resources and allocate network resources.

And this year's popular AI model also puts forward new requirements for the cloud network. Sun Chenghao told titanium media that from the current industry development, the network is still a "constraint" hindering the development of the AI model, "although we are now in the network technology relative to the AI application requirements to do multi-path capabilities, as well as the ability to follow the road detection, faster diagnosis and detection of the short board inside the network to do recovery. But I think there is still a certain gap for the really large-scale training and the construction of AI clusters, which is still a big obstacle to the development of the whole AI application. " Sun Chenghao said.

Looking forward to the future, under the wave of enterprise digitization, the network as the infrastructure of the digital economy, the future network must be flexible, reliable, convenient and high-performance, and with the development of the AI model, on the one hand, it puts forward higher requirements for the cloud network; on the other hand, under the background of this AI model reconstructing all industries, the AI model will also empower the cloud network and make the cloud network better serve the end users.

Welcome to subscribe "Shulou Technology Information " to get latest news, interesting things and hot topics in the IT industry, and controls the hottest and latest Internet news, technology news and IT industry trends.