Why does the domestic cloud computing industry lag behind? Who can shake Alibaba cloud's position? Why is the market competition becoming white hot?

In the history of cloud computing industry, a dialogue between several Internet giants cannot be avoided.

In 2010, the "IT leaders summit" scene, in the discussion of cloud computing, Robin Li "impolitely" said that cloud computing is "old wine in new bottles", and Ma Huateng thought that cloud computing became "water and electricity coal" to wait for hundreds of years or even one thousand years.

Only Ma Yun is different. He said that if Ali didn't do it, he would die in the future.

At this time, Alibaba cloud has been doing it quietly for two years. Wang Jian, specially recruited by Ma Yun, put forward the general direction of "going to IOE" to Alibaba and started the self-study of the cloud operating system "Feitian". Some Alibaba people don't believe it. They satirize Wang Jian, "a doctor of psychology can become the CTO of Alibaba. It's good to learn psychology!"

Just like any new thing is bound to go through the stage of being verified, cloud computing has gone through such a road since its inception in China in 2007: at first, only a few people were optimistic. They walked the first few kilometers alone and were widely questioned. Later, the road became wider and wider, and more and more players went with them, catching up with each other in the crowded track.

This "going to IOE" road was completed in 2013. It represents Alibaba cloud's phased success and ignites the domestic cloud computing market. Since then, players from all parties have accelerated to enter this track to make up for past strategic misjudgments and catch up with Alibaba.

So far, the battle of cloud computing has been fought for nearly ten years, and the victory or defeat has not been decided. However, in the midst of stumbles, the domestic cloud computing industry is making great strides to continue to catch up with Europe and the United States.

Catch up: five years behind

In February 2002, Bezos asked Colin Blair and Frederick, the heads of the United Company, to develop an API that allows developers to access Amazon. This API is the original form of AWS.

If you count from this year, Alibaba cloud's start is five years later than Amazon. In fact, the domestic cloud computing industry as a whole also lags behind Europe and the United States for about five years.

In China, Alibaba was the first to invest heavily in cloud computing, starting in 2007. For Alibaba at that time, it was a helpless move.

In 2008, the annual trading volume of Taobao was 99.96 billion yuan, more than double the 43.3 billion yuan of the previous year. However, at this key node of business growth, Ali's hidden dangers in the underlying it architecture began to be exposed:

The database is too large and there is almost no room for further expansion, and the purchase, operation and maintenance cost of IOE architecture is too high.

Wang Jian once calculated an account for Ali. He said that according to the purchase speed at that time, buying machines and software alone was enough to bankrupt Ali.

Ali chose to "go to IOE", that is, replace minicomputers with cheap PC servers, replace Oracle with open-source self-developed databases, and no longer use high-end storage devices.

This year, cloud computing has not been widely recognized in China. According to Gartner, the domestic cloud computing market is only US $353 million, accounting for only about 1% of the global cloud computing market. In contrast, during the same period, foreign cloud computing giants such as Amazon and salesforce began to release their own PAAS platforms.

The lag of domestic cloud computing industry is determined by the following three factors:

First, in all previous IT technology revolutions, the leading forces are overseas IT giants, and their position is difficult to shake.

It products have two characteristics: high universality and high coupling. The former means that the IT industry, especially software products, are not limited by physical factors and can achieve universality. The latter means that the upstream and downstream of IT products have high interdependence. Taking windows as an example, the application software and hardware coupled with it are difficult to be compatible with the emerging operating system, so its position is difficult to shake.

This also led to the fact that for a long time, most of the IT architectures of domestic Internet enterprises used IBM minicomputers, Oracle databases and EMC centralized storage. This IOE architecture dominated by IBM, Oracle and EMC was also the standard configuration of most enterprises at that time.

Second, most of the growth models of early Internet enterprises refer to overseas, and new models and new industries often start in developed countries first.

For example, Alibaba launched Taobao only the second year after Amazon began to make API.

Third, before cloud computing is proved, it itself faces value disputes.

Cloud computing itself is not a new network technology. Its core is to realize the large-scale allocation of resources. Users can configure the underlying physical resources such as computing, storage and network flexibly on demand and pay according to the usage.

For example, in the traditional mode, an enterprise needs to purchase underlying resources such as servers, databases and storage. These resources are more hardware configuration in the early stage, with high deployment cost and expensive operation and maintenance costs.

Moreover, it is difficult for enterprises to predict the demand when deploying it resources. In order to meet the demand, it is easy to "buy more but not less", resulting in resource redundancy.

On this basis, a superficial understanding of cloud computing is actually one or several enterprises that collect resources and lease these infrastructure to customers in different ways.

Therefore, the technical barrier of cloud computing lies in how to realize more effective resource allocation and more complete service experience in the cloud.

But the biggest miracle of China's Internet is that most players are not industrial pioneers, but their growth rate is often the fastest.

Soon, some large companies in China have accelerated the layout of cloud computing:

Huawei - Huawei cloud was established in 2011 and released the cloud sail plan 2012. In 2015, it released the "Huawei Enterprise Cloud", a public cloud platform, and upgraded the cloud strategy in 2016;

Baidu, Baidu's first opening up to the outside world in 2012, began to accelerate in 2016. Robin Li, the former executive of SAP (enterprise application software provider), worked as a general manager of Baidu cloud, and expanded the team size from 200 to 1100 in one year.

Tencent - in 2013, Tencent cloud was established as an independent brand, Tencent cloud computing company was established in 2014, and Tencent cloud became a strategic business in 2016;

Their official end means the real arrival of the cloud computing era and the official opening of the cloud war.

02 path: Manufacturer's multiple choice questions

The outsiders once ignored Robin Li and Ma Huateng's early disregard of cloud computing as a strategic miscarriage of justice.

The fact is that later, a series of players, including Tencent, Baidu and Huawei, inevitably embarked on the road of catching up with Alibaba cloud.

We also see that around 2016, cloud computing was placed at a strategic high position by more and more Internet giants:

In 2016, Zhang Yaqin, the leader of Baidu cloud, once said that the overall market of cloud computing was a marathon. Ali ran one kilometer and Baidu may have run 0.5 kilometer; In the same year, Tang Daosheng of Tencent said that the cloud business is the position that Tencent must win; Also in this year, Huawei proposed that the enterprise cloud should reach $10 billion in 2020.

Over the same period, Alibaba cloud's first mover advantage has become more obvious: according to the data previously released by Market Research Institute IDC, Alibaba cloud's share in China's public cloud market reached 47.6% in the first half of 2017.

Before discussing the path of giants' layout of cloud computing, let's have a general understanding of the different models of cloud computing and the development trend of the industry:

According to different demand levels, cloud services are divided into IAAs, PAAS and SaaS, respectively corresponding to the rental of underlying infrastructure, software development platform and application software;

The difference between the three is that IAAs layer and PAAS layer have high entry threshold, high market concentration, and SaaS is relatively weak. Internet giants often have massive IT infrastructure and form high capital barriers. Therefore, IAAs layer is mainly giant games, such as Alibaba, Amazon, etc.

At the same time, IAAs layer has more scale effect and is easier to assimilate.

According to different service modes, cloud computing is divided into public cloud and private cloud, as well as later extended "hybrid cloud" and "exclusive cloud".

The core difference between the two lies in the dominance of the data center. The data center of the public cloud is generally provided by the cloud service provider. Under the private cloud service mode, the data center is generally built and operated by the enterprise itself or by the private cloud manufacturer. The hybrid cloud is the different IT systems of the enterprise, which use the private cloud and Public Cloud Architecture respectively.

As shown in the figure above, the advantage of public cloud lies in preferential prices, which is more suitable for small and medium-sized enterprises and innovative enterprises. The advantage of private cloud is mainly reflected in high security, self-control and more suitable for large and medium-sized enterprises that are more sensitive to data security.

Regardless of the classification, the cloud service market reflects two core trends:

First, the IAAs layer focuses on infrastructure, the PAAS layer focuses on platform ecology, and the SaaS layer focuses on applications and more on C-end user scenarios. Anxin Securities pointed out in a research report that among the three, PAAS platform will become the focus of cloud computing ecological game, mainly because SaaS applications are more attached to PAAS platform.

Second, the hybrid cloud model of public cloud + private cloud will be the mainstream service mode in the future.

On this basis, it is not difficult to find that there are actually path differences in the layout of cloud computing among giants:

Amazon AWS and Alibaba cloud represent a bottom-up path from the IAAs layer to the SaaS layer. They all choose to fully develop the cloud computing operating system, typically Alibaba's "flying sky". The core difference between the two is whether the cloud computing ecosystem is completely autonomous and controllable.

In the domestic market, other cloud service providers other than Ali uniformly choose to conduct self-development based on the open source operating system openstack, which is jointly developed and initiated by NASA and Rackspace. The operating system is completely self-developed, which is also an important reason for Alibaba cloud's early difficulties.

Represented by Microsoft, Tencent cloud, baidu cloud, Huawei cloud and Inspur cloud, it is more a top-down path from the SaaS layer.

Microsoft is the most typical representative. It starts from terminal specific services such as office, makes use of the ecology and integration of products, turns to storage, computing and network infrastructure, and then realizes the mutual diversion of SaaS services with IAAs and PAAS.

From the perspective of service mode, whether to choose "public private" or "private public" depends on the positioning and role of cloud service providers.

At the initial stage of market development, the cloud computing market relies more on customer resources. On this basis, Huawei cloud, telecom Tianyi cloud and Alibaba cloud actually represent three forces:

1. The advantage of traditional ICT enterprises lies in huge enterprise resources;

2. The third-party IDC enterprise has the advantage of government resources;

3. The advantage of Internet giant enterprises lies in the enterprise resources in the ecosystem;

We can see that Huawei cloud started as a private cloud and turned to a public cloud in 2015; Whether Alibaba cloud or Tencent cloud, most of their initial businesses first provide public cloud services for enterprises in their own ecosystem.

03 competition: competing for the fourth cloud

From the current market pattern, Alibaba cloud belongs to the first echelon, while Tencent cloud and Huawei cloud are in the second echelon.

According to the data of China's public cloud market in the first quarter of 2021 released by IDC, the domestic IAAs + PAAS market reached US $4.632 billion in the quarter, of which Ali accounted for 40% of the market share, followed by Tencent cloud and huaweiyun, accounting for 11%. Other markets are carved up by Tianyi cloud, Baidu intelligent cloud, Jin Shanyun, wave cloud and other manufacturers.

From this point of view, the current industry competition has two uncertainties:

1. Can Tencent cloud and Huawei cloud shake Alibaba cloud's position?

2. Among the other players, who will be the first to become the fourth stable player?

From the perspective of the industry, after brewing and exploration in the past few years, the main players in the cloud computing market have basically completed the ecological construction of the bottom layer. The next competitive focus will be more on the market scale.

The development of public cloud has a self reinforcing effect: it forms a first mover advantage with capital investment, opens customer acquisition channels with cost performance, and improves profitability with value-added services. In the past, the growth paths of Amazon AWS, Microsoft azure and Alibaba cloud all reflected the above commonalities. Therefore, the public cloud platform market is likely to reflect the trend of the stronger.

Since taking the lead in realizing large-scale, Alibaba cloud has reduced prices more than 10 times a year since 2014, with an annual decrease of up to 30%.

Even so, Alibaba cloud is still the only profitable cloud manufacturer in China. In the past three quarters, its adjusted EBITDA was 24 million yuan, 308 million yuan and 340 million yuan respectively.

Therefore, it is even more difficult for Huawei cloud and Tencent cloud to catch up with Alibaba cloud. But we also see that the competition between them is evolving in the direction of white hot.

For example, as early as March 2017, the bid winning announcement of Xiamen government affairs network cloud service showed that Tencent cloud won the bid at the price of 0.01 yuan. A picture exposed online shows that among the other four enterprises participating in the bidding, the lowest bidding price is 1.7 million and the highest is 3.09 million.

For another example, Alibaba cloud and Tencent cloud began to introduce a hierarchical agent system in 2019. As of August this year, Tencent cloud has more than 8000 channel partners, Alibaba cloud has 10000, and Huawei cloud has exceeded 10000 in May last year.

This is an epitome of the current battle in the cloud computing market:

First, government and enterprise customers are the new focus of competition.

IDC data show that in 2020, the scale of China's Government cloud public cloud market reached 8.14 billion yuan, a year-on-year increase of 61.59%. In this sub market, the top four manufacturers are: China Telecom Tianyi cloud, Alibaba cloud, Huawei cloud and Inspur cloud, accounting for 81.7% of the market.

Previously, the management of Huawei's enterprise business and Alibaba cloud told the media that the government enterprise market is a key breakthrough in 2021.

Second, the industry is likely to start a price war and a resource war;

A cloud computing agent once told the media, "Tianyi cloud's products are basically sold to closely connected government enterprises or Telecom's own businesses, usually packaged and sold at cost."

Tal previously moved from Alibaba cloud to Tencent cloud, and Tencent provided advertising resources for a period of time; Pinduoduo adds Baidu cloud. An important consideration is that Baidu gives it mobile traffic.

It can be predicted that the sale mode of price reduction and resource packaging is likely to become the norm of manufacturers' game.

Third, the market hand to hand battle has begun;

The competition of agents among different cloud manufacturers is an incision for us to observe this war.

Photon planet pointed out in an article that in order to compete for customers, agents even return the Commission points given by manufacturers to customers according to their own highest bottom line. Some agents said, "some agents can give all the Commission points given by manufacturers to customers in order to grab customers. They would rather not make money for themselves than renew later."

The top three are hot, and other players are also fighting in the remaining market plates.

Some of them have surrendered. For example, meituan closed the public cloud service last March; Some have just entered this market. For example, at the beginning of this month, the volcanic engine was upgraded to byte enterprise level technical service business segment and officially entered the public cloud market.

They all want to be the "fourth cloud", but even if they become the fourth, their days are not necessarily