Factors Affecting Cloud Server Network Speed

To improve a specific metric, it's essential first to identify the factors affecting it. The network speed of a cloud server depends on three factors: client bandwidth, intermediate route bandwidth, and server bandwidth. Any of these factors can become a bottleneck. Just as the time it takes to commute from home to the office depends on the congestion at your doorstep, along the route, and at the office entrance, any blockage along these segments affects your commute time. Thus, the network speed of a cloud server is determined by the smallest of these three factors.

1.Client Bandwidth

Client bandwidth refers to the network bandwidth of endpoint devices such as computers or smartphones. With gigabit bandwidth and 5G networks becoming prevalent, client bandwidth is usually abundant and, therefore, not often a bottleneck for network speed. This does not apply if the client itself is a server, in which case server bandwidth analysis applies.

2.Intermediate Route Bandwidth

Intermediate route bandwidth refers to the network bandwidth of the routing nodes that the connection passes through from the client to the server. The route typically consists of multiple routing nodes forming several network channels. For example, accessing a server in London, UK, from a third-tier city in China may involve routing through provincial capitals and international gateway cities like Beijing, Shanghai, Guangzhou, detouring through the United States before reaching the UK. Due to China's insufficient international export bandwidth, the bottleneck for accessing overseas servers often lies in the intermediate route bandwidth. Experience shows that Hong Kong and the western United States offer the best bandwidth for accessing servers from China, followed by South Korea and Japan, and then Europe, Southeast Asia, the Middle East, Africa, and Latin America. Generally, the less developed the local internet infrastructure, the smaller the international export bandwidth from China.

3.Server Bandwidth

Server bandwidth refers to the network bandwidth of cloud or dedicated servers, set by the cloud computing platform or data center. For a specific access request, the actual server bandwidth used depends on the number of connections requesting the server simultaneously. The more simultaneous connections, the smaller the bandwidth allocated to each connection. Network protocols allocate a portion of the total server bandwidth to a specific access request based on bandwidth distribution mechanisms.

In summary, cloud server network speed depends on client bandwidth, intermediate route bandwidth, and server bandwidth, determined by the smallest of these three factors. When using cloud servers, the product parameters only include server bandwidth, with intermediate route bandwidth, mainly international export bandwidth, often being the reason for slow speeds on overseas cloud servers. Therefore, when analyzing cloud server network speed, it is crucial to consider both intermediate route bandwidth and server bandwidth, as focusing solely on server bandwidth values is meaningless.

General Principles for Selecting Cloud Server Bandwidth

Given the variety of use cases for cloud servers and the complex factors affecting their speed, specific issues require specific analyses. However, a few general principles can guide the selection process.

1.Identify the primary direction of traffic for the cloud server, whether it is outgoing from or incoming to the cloud server. For incoming traffic, such as accessing websites or downloading software remotely to the cloud server, server bandwidth considerations are unnecessary, as incoming bandwidth is unlimited. This principle is also one basis for DDoS attacks. If network speed is unsatisfactory in this scenario, the bottleneck is likely due to intermediate route bandwidth.

2.Intermediate route bandwidth, especially international export bandwidth, rather than the cloud server's own bandwidth, often constitutes the biggest bottleneck for accessing overseas servers from within China.

3.Intermediate route bandwidth is determined by service providers and international gateway conditions and is beyond control. The need to navigate international links and service provider routing nodes when connecting from mainland China to overseas servers can lead to instability due to international link congestion and service provider routing restrictions. Different service providers may have different routing nodes for accessing overseas servers, leading to varying experiences among users. The reality of insufficient and uncontrollable international export bandwidth must be accepted if the client and server are in different countries or regions. Improvements can be made through intermediary nodes, CDN acceleration, or dedicated lines, but these solutions require higher technical and financial investments.

4.Determining whether server bandwidth is a bottleneck can be done by analyzing outgoing bandwidth in monitoring data. If the bandwidth limit is reached, upgrading server bandwidth or using multiple cloud servers to distribute traffic may be necessary. If the bandwidth limit is not reached, server bandwidth is likely not the bottleneck.

Other Issues Regarding Network Bandwidth

After understanding the factors affecting cloud server network speed and the general principles for selecting bandwidth, exploring related concepts becomes meaningful. Without this foundation, delving into these concepts might lead to confusion and distract from the goal of improving network speed.

1. What are Dedicated and Shared Bandwidth?

Cloud server specifications with specific bandwidth values typically indicate dedicated bandwidth. Dedicated bandwidth means each cloud server has a set bandwidth limit it cannot exceed. However, all cloud servers still share the bandwidth of the entire cloud computing platform or cabinet. If the bandwidth limit is set too high or the platform's total bandwidth is too low, bandwidth resource contention may still occur.

Cloud server specifications that list unlimited bandwidth refer to shared bandwidth. Shared bandwidth means no specific bandwidth limit is set for each cloud server. The number of servers sharing the bandwidth and the total bandwidth available are determined by the cloud server provider and may adjust with business growth, usually not disclosed publicly.

Thus, compared to shared bandwidth, the chance of bandwidth contention among cloud servers is lower with dedicated bandwidth. However, dedicated bandwidth does not guarantee bandwidth usage entirely. Dedicated bandwidth lacks the flexibility of shared bandwidth, which can accommodate higher network peaks. Dedicated and shared bandwidth have their respective advantages and disadvantages in terms of stability and speed, with no clear superiority.

Choosing between dedicated and shared bandwidth may seem challenging. However, for downstream bandwidth, such as uploading files to the server, the bandwidth is always unlimited. Moreover, for overseas cloud servers, the bottleneck typically lies in intermediate route bandwidth, with the server's own bandwidth usually being sufficient.

2. What are Upstream and Downstream Bandwidth?

The concepts of upstream and downstream bandwidth, as well as inbound and outbound bandwidth, can be confusing. The following table clarifies the differences between these concepts.

3. How are Bandwidth and Traffic Converted?

For example, a cloud server with 10M bandwidth running at full capacity 24/7 for a month results in a total monthly traffic of 30x24x3600x10Mb/8=3240000MB, or approximately 3.1TB. Thus, 10M bandwidth fully utilized equates to a monthly traffic of 3T, and 1M bandwidth fully utilized equates to 300G. Note that bandwidth is measured in Mb/s, while traffic is measured in GB or TB. Since 1 byte equals 8 bits (1B=8b), the calculation formula requires division by 8.

4. What is the CN2 Route?

The CN2 route, officially known as China Telecom's Next Generation Carrying Network, offers more advanced technology, less congestion, and faster connection speeds compared to the 163 route. Currently, only a few countries, including Hong Kong, Taiwan, Japan, the United States, and South Korea, have some cloud server providers connected to the CN2 route. Additionally, the CN2 route costs more than ordinary routes.

Using the example of a Hong Kong CN2 cloud server from Zhaomu Data, a tracert route from a local computer shows provincial, international, and backbone routing nodes with IP addresses starting with 59.43, indicating the highest grade CN2 GIA route (see lines 6-10).

C:\Users\zhaomu.com>tracert 43.224.153.136
  1     1 ms     1 ms     1 ms  192.168.0.1
  2     3 ms     5 ms     6 ms  192.168.1.1
  3     7 ms     6 ms     5 ms  124.74.56.214
  4     9 ms     5 ms     *     101.95.91.189
  5     *        *        *     请求超时。
  6     *       10 ms     *     59.43.80.146
  7     7 ms     *        7 ms  59.43.138.62
  8     9 ms    10 ms     6 ms  59.43.187.74
  9    31 ms    38 ms    38 ms  59.43.188.230
 10    40 ms    37 ms    37 ms  59.43.250.26
 11    40 ms    35 ms    34 ms  118.184.22.10
 12    35 ms    35 ms    35 ms  69-172-92-253.static.imsbiz.com [69.172.92.253]
 13     *        *        *     request timeout。
 14     *        *        *     request timeout。
 15    33 ms    31 ms    37 ms  69-172-92-086.static.imsbiz.com [69.172.92.86]
 16    62 ms    40 ms    37 ms  43.249.3.6
 17    32 ms    34 ms    33 ms  43.224.153.136

This article analyzes factors affecting cloud server network speed and provides general principles for selecting cloud server bandwidth, aiming to assist readers in choosing suitable cloud servers.