Shulou(Shulou.com)11/24 Report--

As the community of start-ups, growth companies and individual developers grows, so does their need for an easy-to-use, energy-efficient cloud infrastructure. The scale of these group businesses is often limited to small-scale use or only used as an internal development environment, mainly used in e-commerce store management, enterprise websites, Mini Program operation and other scenarios. At this time, changing to a cloud server that is lightweight, efficient, easy to deploy, affordable, easy to use and easy to manage is naturally the first choice for this group.

Huawei Cloud has launched a new cloud cloud server L instance (hereinafter referred to as "HECS L") for start-up and growth enterprises and starter developers with relatively weak IT capabilities. compared with traditional cloud servers, HECS L integrates cloud servers, network, storage and security, further simplifies parameters and options, and provides a variety of application images and system images to facilitate rapid business deployment. At the same time, under the premise of ensuring performance and security, the price of HECS L is also very affordable. For start-up and growing enterprises and individuals, they can deploy, manage and use resources through visual pages without much knowledge of operation and maintenance.

As the latest product of Huawei Cloud, HECS L is based on Huawei Cloud Entrepreneurship architecture, which consists of two parts: the data side "soft-hardware collaborative system" and the management surface "Yaoguang Smart Cloud brain". The data surface is vertically optimized for L0 / L1 / L2, resulting in a 2-fold increase in computing density and a 4-fold increase in delivery efficiency. High-power CPU heat dissipation and reduced PUE are realized by liquid cooling. Through the deep cooperation of software and hardware, the "Optimus Card" can realize full unloading and acceleration of computing, storage, network and security, and provide storage delay as low as 100 μ s and network delay of 10 μ s.

As a distributed cloud operating system, Yaoguang Smart Cloud, the management and control surface of the giant architecture, can achieve global resource scheduling, dynamic negotiation and governance, and intelligent matching of diverse computing power, so that the supply of global resources is extremely optimal. the use of diverse computing power is extremely simple. Huawei Yaoguang scheduling algorithm team and Huazhong University of Science and Technology team won two championships, one runner-up and two runners-up in the 12th DIMACS Challenge. In addition, Huawei Yaoguang scheduling algorithm team also won the double track championship in GECCO 2020 OCP&USCP and set more than 50 world best PDPTW records. The strong algorithm team support also brings the strong scheduling capability of Huawei cloud giant architecture.

At the network level, Huawei Cloud Global savings Network KooVerse covers the global distributed cloud infrastructure with a unified architecture. This network provides a high-quality network with security, stability and low latency, and creates a high-quality experience circle for 50 millisecond users to ensure low-latency experiences such as real-time audio and video, real-time games and online education.

Through Huawei Cloud Architecture and Global savings Network, HECS L can achieve efficient service deployment and network coverage, which provides a basic guarantee for an efficient and agile user experience. Next, let's take a look at the performance of HECS L.

Service selection and purchase of HECS L is easier to purchase and use than traditional CVM. Huawei Cloud user experience team adjusts the selection process of HECS L to the pain points encountered by small and medium-sized enterprises and developers in the process of deployment and operation and maintenance, and adjusts the matching mode to the selection and purchase mode based on the customer's actual application scenario. Here, Huawei Cloud provides free application images and system images for users to choose from.

The application image is mainly aimed at customers with special needs, including Linux visualization pagoda panel, WordPress enterprise site building system, Matomo website statistical analysis platform, Portainer Docker visualization, GitLab one-stop DevOps platform, PrestaShop e-commerce system, Nextcloud enterprise cloud disk system, Odoo enterprise integration ERP system, Superset data exploration and visualization platform, SRS real-time video server. These application images cover server operation and maintenance, enterprise site building, data analysis, Docker container, code management, e-commerce, enterprise cloud disk, ERP, data governance, streaming media push and other services, which can basically meet the business needs of most start-up and growing enterprises and individual developers.

In addition, four types of system images, Windows Server, Ubuntu, CentOS and Huawei Cloud EulerOS, are provided to users who need a pure environment for business deployment. The deployment of these four system images is more flexible and suitable for professional developers.

In terms of specifications, compared with other CVM products, HECS L currently provides products from Beijing, Shanghai and Guangzhou for users to choose from. Processor and memory specifications include 2U2G, 2U4G and 2U8G. Different sizes of system disks, peak bandwidth and traffic packets are provided by default for different business scenarios to help users quickly locate their needs. These instance specifications are different from previous server products, similar to the collection of multiple services, including system disk, IP, bandwidth, traffic and other services, among which the most cost-effective is undoubtedly the traffic package mode, which can not only give users more bandwidth, but also reduce costs through the traffic mode. Even after using the current month's traffic, the excess traffic is charged on demand, with a unit price of 0.80 yuan / GB, which will not affect the operation of the business. Considering the start-up and growth enterprises and individual users who choose HECS L, the business scale is relatively small, so the traffic packets included in the example can generally meet the daily use.

In addition, HECS L also provides additional data disk, host security and cloud backup services for specific needs. From the point of view of the purchase method, HECS L basically includes the minimum environment needed for the operation of the CVM, allowing users to build the environment with one click, reduce the decision-making process, automatically orchestrate, activate with one click, and go online in minutes. At present, HECS L offers a prepaid model with packages ranging from 1 month to 3 years. A maximum of 15 units can be purchased in Beijing, Shanghai and Guangzhou. But if you have a need for a larger quota, you can buy it by applying for a quota. According to our experience, Huawei Cloud usually responds to the request within 1 working day.

In addition to the examples of user-selected HECS L-related specifications, Huawei Cloud also offers two packages on the basis of HECS L. the default specification of the basic package is 2U4Gcenter 70G system disk, and provides peak 5Mbps bandwidth, 1000GB traffic, 100GB cloud backup and host security services, which are suitable for startups or personal basic site needs.

In addition, there are two highly available packages of HECS L and two 2U4G hosts. On the basis of cloud backup and host security, cloud cloud load balancing service is also included, which is suitable for start-up and growing enterprises with relatively large business volume and certain technical capabilities, with high performance-to-price ratio, stability and security.

For traditional cloud services, after the purchase of products, resource management is often scattered in multiple user interfaces, which leads to users can not clearly feel which resources they have purchased, and do not know the status of the current resources. HECS L provides visual resource view and information aggregation view, through charts and relationships between resources to help users understand the relationship between resources and real-time dynamics, to achieve global monitoring, unified life cycle management.

Deployment experience next, we will experience the efficient and convenient operation, maintenance, management and use experience of HECS L through the deployment of actual business. Here we use the pagoda panel to build the station. Pagoda panels provide commonly used site-building templates. Huawei Cloud will automatically deploy pagoda Linux panels after the selection is completed.

Through step-by-step instructions, you can help users to reset server passwords, set security groups, and provide management backend links to help users quickly bind domain names.

In this process, users do not need to learn the relevant knowledge of the command line, just wait for the purchase, log in to the management page and complete the simple configuration to enter the pagoda Linux panel and quickly enter the site building panel. Here the pagoda Linux panel provides two common combinations of LNMP and LAMP.

In addition, the pagoda Linux panel also provides a variety of common site building templates, including WordPress, ThinkPHP, Discuz, etc., users only need to choose according to their own business model to achieve rapid site construction. There are also Java, Node, Go, Python and other projects to facilitate rapid deployment.

Here, after installing the PHP, MySQL, Nginx and other running environments, we can use our own WordPress to build the site. From download to deployment and installation, we can quickly complete the website environment in a few minutes.

In the background of WordPress, we can also customize, edit and upload related modules of the website. From the purchase of the server to the deployment of WordPress, we did not enter any code or script during this period, and the whole process has a complete graphical interface, which is very friendly for startups and operators of start-up and growth enterprises.

In addition to using the pagoda Linux panel to build the WordPress site template, HECS L also provides WordPress images. You can quickly build the WordPress site building system by selecting, purchasing, and setting the administrator's username and password. This process does not require users to manually install Nginx, PHP environment and MySQL database, and can be deployed with one click without the command line, which is more efficient.

For e-commerce websites, HECS L also provides PrestaShop application images to facilitate rapid site management, and the whole deployment process is basically the same as the pagoda Linux panel. PrestaShop is very suitable for foreign trade enterprises to set up independent sites to facilitate users at home and abroad to choose and purchase goods.

In addition, the PrestaShop background also provides a sound management environment, including commodity management, order management, customer information management, multi-language and so on, for foreign trade enterprises to quickly build their own sales network.

Generally speaking, HECS L is basically a "fool" operation in the deployment of application images. Users can neither code programming nor understand technical details. They only need to follow the steps of Huawei Cloud console to achieve zero-based business deployment, greatly reducing the threshold for using CVM, using it out of the box, minute-level deployment, so that the business can run quickly from zero, and the threshold for smooth use can be achieved. This is a different experience from HECS L.

Performance testing next, we use testing tools to verify the performance of HECS L. The configuration we use here is 2U4G, mainly for performance testing, and the test environment is Ubuntu. The test items include Geekbench 6 processor performance test, Stream memory test, Sysbench MySQL performance test, ClickHouse benchmark test and Rally Elasticsearch benchmark test. In order to better compare the performance of HECS L with Sysbench, ClickHouse and Rally tests, we also conduct the same test on another lightweight server An owned by another domestic mainstream cloud server manufacturer. The relevant specifications are consistent with HECS L.

Performance Test of Geekbench 6 processor

In terms of theoretical performance of the processor, we use Geekbench 6 to run HECS L.

The average score of three times is 1184 for single thread and 2132 for multi-thread, which is basically consistent with its lightweight and entry positioning.

Stream memory test

Next we use the stream memory testing tool to test the memory performance of HECS and ALS6. Stream is one of the most popular comprehensive memory bandwidth performance measurement tools in the industry. Stream supports four operations: Copy, Scale, Add and Triad.

Copy it first accesses one memory unit to read out the value, and then writes the value to another memory unit. The Scale operation first reads the value from the memory unit, makes a multiplication operation, and then writes the result to another memory unit. The Add operation first reads two values from the memory unit, does the addition operation, and then writes the result to another memory unit. The representation of Triad combines Copy, Scale, and Add operations for testing.

The average score is as follows:

Testing the data here, we see that the HECS L memory performance level matches the processor's processing power exactly. It can meet the deployment of lightweight business in the cloud. Next, let's take a look at how HECS L performs through a few benchmarks that are skewed towards actual use.

Sysbench-MySQL performance test

As a widely used relational database, MySQL is widely used by users all over the world. MySQL software adopts the dual licensing policy, which is divided into community version and commercial version. Because of its small size, high speed and low total cost of ownership, especially open source, the development of small and medium-sized and large websites generally choose MySQL as the website database.

Here we use Sysbench to test HECS L for MySQL read-only, write-only, read-write. The relevant test data are as follows:

Judging from the data, HECS L outperforms A lightweight server with similar location in the read-write, read-only and write-only tests of MySQL. For website construction, cross-border e-commerce multi-store management, enterprise application deployment, Mini Program and Mini Game deployment, which are common in start-up and growth enterprises, this performance matches the business scenario very well.

Clickhouse benchmark test

ClickHouse is an open source column-oriented database management system, which can generate analysis data reports in real time. It is the fastest and most resource efficient open source database for real-time application and analysis. As a very popular database for OLAP, ClickHouse can provide seconds of big data query. With its excellent performance, ClickHouse is widely used in major Internet companies. The data generated by this test is about 1G, and each statement is executed for 30 seconds. The execution data are as follows:

Judging from the data, HECS L performs well in OLAP scenarios. The performance of some scenarios is much higher than that of A lightweight servers, and the average performance is about 3-4 times that of A lightweight servers. Through the test of ClickHouse, we can see that HECS L can well deal with a large number of data queries in a short time, and can meet the business and daily development needs of start-up and growth enterprises.

Rally-Elasticsearch performance test

Elasticsearch is a distributed, RESTful-style search and data analysis engine that addresses the emerging use cases. As the core of Elastic Stack, Elasticsearch stores users' data, allowing users to quickly complete searches, fine-tune correlations, conduct powerful analysis, and easily scale. Elasticsearch can implement log monitoring, infrastructure monitoring, APM, synthesis monitoring, enterprise search, Maps, SIEM, terminal security and other scenarios. Here, we test the performance of Elasticsearch through Rally. The test results are as follows:

In the relevant test scenarios, the throughput of HECS L reaches 35318.52 docs / s. Compared with A lightweight server products, HECS L performs very well, almost twice the throughput of A lightweight server products.

Through the benchmark test, we can see that many application scenarios of HECS L are relatively balanced. Under the similar price and hardware configuration, compared with other lightweight server products, the performance of HECS L is undoubtedly very outstanding. The matching users and needs of this product are consistent. Combined with the attractive price and convenient deployment mode, it can be said to achieve "three steps to build a station, twice the performance." The first step in simply going to the cloud.

Recently, Huawei Cloud 828 marketing season has been launched. Cloud Yao Cloud server L instance 2 core 2G 3M costs only 29 yuan 3 months after receiving the coupon, and 2 core 4G 5M costs only 228 yuan a year. For start-up and growth enterprises and individual developers, such products are of course a cost-effective choice.

Activity address: click here to enter

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.