Photo Source @Panoramic Network
Text: Guo Hua
Looking back on the past, after years of development, the market structure of cloud computing has been basically determined. Over the past few years, the cumulative expenditure of AWS, Microsoft and Google, the three giants of cloud computing, on capital such as data centers, has exceeded $200 billion, and is increasing. Now that the product form has been basically determined, every sum of money of the giant is a brick on the threshold, making it more and more difficult for new players to enter.
Unless the product form of cloud computing changes, it will be very difficult to change the market structure.
But IBM said he found an opportunity.
In 2018, IBM purchased Red Hat for $34 billion, declared itself the number one provider of hybrid clouds, and said it would radically change the cloud computing landscape.
It's the biggest acquisition in open source history, and even in the global technology industry, it's in the top three.
At that time, Red Hat's share price was $116 per share, while IBM's price was $190, with a premium of more than 60%. The whole deal is done in cash, which is not easy for IBM with a market capitalization of just over $100 billion. It will not only lead to high liabilities, but also leave a legacy that will be difficult to recover in the short term. In fact, for this acquisition, IBM had to suspend its stock repurchase program, which had been going on for many years, as his main way of stimulating stock prices in the past.
IBM is desperate, and what supports his decision is his good expectation of hybrid clouds.
According to his calculations, the hybrid cloud has a market size of up to $trillion, and the acquisition of the red cap has made IBM the number one supplier of hybrid clouds. In contrast, the global cloud market in 2018 is about $8 billion, of which AWS, the first one, is about one-third, roughly $250 billion.
This is very worthy of discussion.
Because of the calibre problem, the market size of hybrid clouds has not been accurately described. IBM says there is a trillion, while Statista says there is only 90 billion. In order to avoid drooling, I will try to ignore this number and focus on the logic of the problem itself, that is, whether IBM can change the cloud computing pattern through hybrid clouds.
01 What is a hybrid cloud?
Hybrid cloud is a form of hybrid use of public and private clouds.
Generally speaking, customers will put the core or stable business into private sector, edge or flexible business into public cloud, while opening up data and applications in both environments, unified management, and adjust workload between public cloud and private cloud as business changes.
Hybrid clouds are not strictly defined and sometimes include cloudiness, i.e. mixing resources from multiple public cloud vendors to eliminate customer lock-in concerns. Cloudiness is outside the scope of today's discussion.
02 Where do mixed clouds come from?
Internet is the carrier of information civilization and the training ground and origin of new technologies. It is not so much that cloud manufacturers have Internet genes, but rather that the Internet has spawned cloud computing. There are two meanings of this boom: one is the growth of cloud vendors in Internet companies, the other is that early customers of cloud computing are also Internet entrepreneurs. The reason for this is that the scenarios are similar. Cloud vendors'products often come from successful practices in their Internet business. Therefore, they are universal in similar scenarios and are easier to export in the form of cloud products. Whether it's AWS abroad or Aliyun in China, it's better.
However, with the rise of Internet entrepreneurship, cloud computing will also follow him.
Put aside the problem of the base number, the main reason is that the dividend of Internet entrepreneurship tide is almost eaten, and those who need to go to the cloud are also on. Many of the remaining IT market are traditional corporate customers. Compared with the Internet, traditional enterprises have two characteristics. One is that they have been established for a long time. Many of them have their own computer rooms after long-term construction. The other is that their business is relatively stable. The flexibility of cloud computing is not so urgent. Moreover, some industries, such as finance, still have restrictions on industry supervision. Law uses public clouds.
Hybrid clouds are generated in this context.
Through hybrid cloud, while maintaining their own computer room and core business, traditional enterprises can put some marginal business or business with high elasticity requirements on the public cloud, and upgrade the internal technology architecture with the help of cloud manufacturers, so as to make reserves for the future, both fish and bear's paw.
03 Where is the Mixed Cloud going?
Next, let's see where the hybrid cloud is going.
From the abstract point of view, hybrid cloud is actually born in a contradiction. It is the product of the game between cloud vendors and customers in terms of business pressure, historical accumulation, technology, cost and control power. Where he wants to go depends on how this contradiction develops.
For traditional enterprises, firstly, the core IT business of traditional enterprises is relatively stable, and they often use the products of traditional hardware and software suppliers, such as IBM minicomputers, Oracle databases, etc., that is to say, to go to the storage of IBM machines, Oralce databases and EMC mentioned in IOE. This set of things is not only expensive, but also able to cope with it. Pay for daily work, so there is no need to move; Second, in the trend, the enterprise's IT construction actually abides by the rule of informatization of the core business to the Internet of some business and then to the overall digitalization. The former software suppliers and their products are not good at the Internet and digital construction, so the traditional enterprises. Third, the core business of traditional enterprises is not IT, so the IT talent reserve is insufficient and self-construction is unrealistic, which requires the introduction of external forces. Fourth, at the same time, the habits and business attributes of traditional enterprises also determine that they have higher requirements in terms of security and isolation, such as production line correlation. Once the system is broken down because of network security, the result may be catastrophic. There are also some industries, such as finance, which are restricted by special laws and regulations of the state.
For public cloud vendors, firstly, most of them are not software companies except Azure, and their business is flexible and changeable, so the construction of core business systems, such as databases, is often not as good as that of traditional software vendors such as Oracle; secondly, public cloud vendors have basically successful Internet business, in the Internet and in the number of Internet. Characterized area accumulates a large number of products, but also a large number of scenario solutions, which can just make up for the demand gap of traditional enterprises. Third, public cloud vendors rely on big trees besides their own business, and have sufficient cash flow. With the advantages of funds, talents and scenarios, cloud vendors are building at the level of basic software. In Gartner's 2008 database magic quadrant, for example, AWS has entered the leader quadrant with traditional vendors such as Oracle and SAP, and Aliyun and Google have also entered the visionary quadrant. Fourthly, cloud vendors are also investing in enhancing security and isolation. Besides product level, they are also introducing private products. Cloud or industry cloud is gradually meeting the requirements of enterprises and laws and regulations. Fifthly, cloud computing has better flexibility and lower comprehensive cost, which meets the general trend in economics.
In this way, traditional enterprises have historical accumulation and higher restrictions, but there is a large demand gap, which gives cloud vendors opportunities. Although cloud vendors are not dominant in basic software, they are developing rapidly. At the same time, there are a lot of solutions to meet the needs of traditional enterprises. At the same time, in the game between the two, the traditional software suppliers of the traditional enterprises are trying to meet the new demand, just because of the cost structure, the transformation is difficult. This is another problem, which will not be discussed today. In a word, these contradictions determine that the IT environment of traditional enterprises is more complex, including systems developed and deployed by themselves, privatized products of traditional software suppliers, private and public clouds of cloud vendors.
Hybrid clouds are born in this contradiction and will develop with these contradictions.
Enterprise core IT needs will not shrink, but the proportion of Internet and digital transformation will gradually shrink, enterprise isolation requirements will not be reduced, but with the gradual improvement of cloud vendors'products or solutions and gradually not a problem, this growth and decline, the proportion of cloud will become higher and higher, until one day, revenue exceeds the cost of migration. Core IT systems will also be in the cloud, so the hybrid cloud will disappear.
Only this process will be long, because offline still occupies more than 80% of the overall IT market, which is a protracted war.
Does 04 IBM have a chance?
Does IBM have a chance to change the landscape of cloud computing in this protracted war?
So we have to mention two ways of thinking about hybrid clouds.
The first is the idea of public cloud manufacturers, who use the hybrid cloud as a bridge for customers to cloud, extend the public cloud down to the customer's private computer room, and then gradually migrate the workload, with the ultimate goal of all the customers on the cloud. AWS's Outposts,Azure Stack and Google's Anthos are the products of this idea, and I temporarily call this form cloud vendor hybrid cloud.
The second is the thinking of traditional IT vendors, such as IBM and Red Riding Hood, who used to be privatised and deployed, but now that they want to be hybrid clouds, they will naturally move from the more familiar and good offline to the cloud as a supplement to the offline. They don't have a real public cloud themselves, so they have to try.
To discuss whether IBM has a chance, we just need to discuss whether independent hybrid clouds can be dominant.
According to the logic of independent hybrid clouds, they are
This reminds me of a theory called aggregator.
The theory was invented by Ben Thompson, a famous analyst. He said that there are three roles in any business activity: supplier, channel and consumer. There are usually two ways to make a big business. One is horizontal integration, the other is vertical integration. For example, for a channel, he can integrate other channels horizontally, and then control distribution to obtain pricing power upstream and downstream; he can also integrate suppliers vertically, and then reduce costs or improve the quality of products and services through full-link optimization. He also found that before the Internet, successful vertical integration tended to be backward, that is, channel integration of suppliers and then services to customers, such as newspapers, hotels or television stations. But in the Internet era, successful integration often goes forward, that is, channel integration of customers in turn affects suppliers, such as Airbnb, Uber and Google. Le.
Ben Thompson calls this pattern aggregator. Correspondingly, the underlying theory is aggregation theory.
In aggregation theory, aggregator has three typical characteristics. First, it directly establishes links with customers. Second, it has zero marginal cost (digital products, unlimited replication). Third, it gradually reduces the cost of customer acquisition with the expansion of scale. The third point is the most attractive point of the theory of aggregation, which decides that aggregators can win all. For example, Uber, the more customers he has, the more drivers he attracts. The more drivers he has, the more competitive the platform will be. In turn, it will attract more customers. As the platform expands, the cost of getting customers will gradually decrease. The faster he runs, the harder it will be to catch up with his competitors once he runs.
If the independent hybrid cloud fits the aggregation theory, with years of accumulation of Red Hat and IBM in the traditional IT market, they can absolutely gather a wave of start-up customers and roll up the snowball. The key to conformity lies in the third point of aggregation theory. In this point, aggregators must be able to continuously improve their attractiveness through aggregating suppliers, and consequently continuously reduce the cost of customer acquisition, which requires that there are enough suppliers in the market for integration.
For an independent hybrid cloud, its supplier is a public cloud vendor, and that becomes the question of how many public cloud vendors there will be.
You may have heard of the five clouds, but in fact, it stems from a joke told by IBM's old Watson in 1943, when he said that only five computers are enough for the whole world. This is obviously wrong, but in the age of cloud computing, this joke has become the ultimate assertion and is regarded as the standard.
This is because cloud computing is a heavy asset, heavy business. It has been mentioned that the AWS, Microsoft and Google Big three have invested more than $200 billion in the global data center for a long time. If the latecomers want to catch up, they must equalize the investment first, while the starters can fight the price war with the new players without losing or losing money, allowing the new players to pay much more than they originally paid. So with the development of time, the threshold of cloud computing will be higher, and there will be only a few giants left in the market.
And that hasn't taken into account that the existing cloud vendors have grabbed a large number of customers.
So in this case, the value of independent mixed cloud platform is not great, even if the original strong cloud giant lying flat let him docking, he can only support the docking three or five, the information is too transparent, cannot allow middlemen to earn the difference, also cannot build a moat, cannot meet the requirements of the aggregator.
This means that the independent hybrid cloud can not be dominant, which means that IBM can not change the market structure of cloud computing by independent hybrid cloud.
After concluding, I would like to continue discussing some surprises.
In some cases IBM has a chance.
That is, cloud vendors are highly standardized and degenerated into generic IAAS services, on the basis of which independent hybrid clouds build aggregators, aggregating a large number of software vendors to provide PAAS and SAAS,. In this case, the real supplier of aggregators is these software vendors, which is enough. In fact, at first everyone said that cloud computing is like water and electricity, and general IAAS service is the most water and electricity part.
From the description of Openshift, IBM is now the idea, his first supplier is open source software, so IBM calls him the open hybrid cloud.
However, it is unlikely that this assumption will come true. Cloud manufacturers will not be content to do IAAS, alone and will not only provide standard products. At the same time, cloud manufacturers have their own cloud market and are teaming up with a variety of software vendors.
Even if it doesn't count, IBM will still run into Google's hybrid cloud platform head-on.AnthosIBM's Openshift is based on K8S, and Google's Anthos is also based on K8S, which is called open hybrid cloud, Google is called open cloud, and K8S is open source for Google.
However, I would also like to quote a passage from historian Professor Swartz:
As long as the future doesn't happen, everyone has a chance.