Clouds and Powder Keg Open Source

“Europe today looks like a powder cellar, and leaders are like people smoking inside. One spark will cause an explosion that will bury us all. I don’t know when this will happen, but I know where. Everything will be ruined by some stupid event in the Balkans ”- Otto von Bismarck, 1878

One hundred years ago, on November 11, 1918, a truce was signed that put an end to the First World War. The number of lost lives in that war is now hard to imagine. For example, in America, the Vietnam War is rightfully considered a military disaster. Over twenty years of fighting, the United States lost 58,318 fighters. For comparison, only in the first battle on the Marne in 1914 did the Allies lose four times as much. In five days.

Some would say that the horrors of war were impossible to predict. The problem is that at least some of the parties involved were well aware of the consequences. British Foreign Minister Edward Gray, after speaking in parliament in support of the war, is believed to have said: “The lights go out all over Europe. In our lifetime they will no longer light up. ”

Therefore, in subsequent years, historians tried to answer the question: if the consequences were understandable, then how did the July crisis allow  - a series of related events, as a result of which the war was the only possible outcome.

Although extremely complex in detail, the answer is simple. Given the atmosphere and political structures of the time, not one of the participants in the events felt that he had an alternative. One of the most frightening things in studying the causes of war, in fact, is that if you look at the political realities of the time, it’s actually easy to understand the justifications for the actions of each state.

In the end, we will agree that war was actually inevitable. The ease of accepting this truth is truly scary.

---

Three years ago, a venture capital fund gathered a small group of media representatives, vendors and analysts, including us, to discuss the importance of open source in business. After the presentation of his own model, the partner of the venture company introduced a group of executives from open source partner commercial companies. Each of them described in detail how open source replaced proprietary alternatives for clients.

Of course, we agree that the transition of developers to open source across the enterprise is changing the nature of the supply. To some extent, this is the main belief that we have been promoting for many years. Back in 2011, we published an article entitled “Bottom-up implementation: end of supply, as we know it”. But in the proposed model, it was interesting not what she talks about the present, but rather what she cannot say about the future.

The event did not mention cloud services directly. It was said that OSS investors and commercial developers compete with proprietary software. No special attention was paid to Amazon and other suppliers of oversized clouds, they were not even called. A question on this subject was politely rejected.

This is interesting, because at RedMonk at that time, when evaluating commercial open source teams, they asked them to answer the standard simple question: “Who is your competitor?” If they called a proprietary alternative, it suggested that the company was turned into the past. If the answer was a cloud, it could be safely assumed that the startup is looking forward.

As you can see, now this thinking has reached the market. In the last 12–18 months, in fact, a coup has taken place. If before companies did not consider worthy of mentioning cloud providers such as Amazon, Google and Microsoft, now they regard them as a deadly threat. The fear of cloud providers has become so overwhelming that commercial open source vendors often, contrary to the advice of consultants, make strategic decisions that violate open source cultural standards, cause massive and persistent negative public relations, and jeopardize relationships with developers, partners, and customers. In particular, they are increasingly turning to models that blur the boundaries between open source and proprietary software in an attempt to gain the advantages of both worlds, but in the end there is a high probability that they will get the disadvantages of both.

Open source commercial providers have taken these steps, having been informed of the risks in advance. This indicates their assessment of their prospects in a world where massive clouds are increasingly dominating, expanding the range of services. Undoubtedly, such strategic decisions have serious, inevitable negative consequences, but commercial open source suppliers - or at least their investors - consider the lack of action an even more destructive option.

It is interesting to see whether this belief persists after the announcement of Amazon Web Services this week. Here is a summary of the story that led to current events:

• 2010 : Written by Shey Banon almost ten years ago, Elasticsearch is an open source search engine with an authorization license. It turned out to be popular enough to eventually form a commercial organization around it. Elastic NV - originally Elasticsearch BV - went through several rounds of financing totaling more than one hundred million dollars, last October it carried out an IPO, and now it is valued slightly below $ 6 billion.
• 2015 : Five years after the project was launched - presumably at the request of customers - Amazon launched a cloud service called Amazon Elasticsearch Service based on this permit. He directly competed with Elastic NV's commercial offerings, both local and cloud.
• 2018 : partly due to competition with this and other clouds, Elastic NV began to blur the boundaries between its open source offering and proprietary licensed add-ons, in particular x-pack. It is noteworthy that Elastic did not follow in the footsteps of some colleagues, but tried to solve the problem using hybrid licenses, but began to mix open and proprietary source code in one repository, and the builds included this non-free software by default.
• 2019 : Amazon has taken several responses this week. First, with the support of Expedia and Netflix, she introduced what she sees as the “distribution” of Elasticsearch. But it is assumed that in all respects it will function as a fork. Secondly, the project includes open source add-ons that are similar to features for which Elastic NV charges a fee, without putting them in the public domain. Thirdly, as in the case of the original AWS service based on Elasticsearch, the company used the name Elasticsearch for the project.

Given that the former contradictions grew into an open conflict, many questions arise. How did it come to that? Was it inevitable? And the obvious question is: who is to blame?

At least one of these questions is easy to answer. This step has been expected for some time. At least since September, when the Commons Clause license appeared :

Of course, it seems unbelievable that cloud providers everywhere will start deploying and licensing open source software under the Commons Clause license from commercial vendors. In fact, Commons Clause can lead to the opposite result. It increases the likelihood that cloud providers will try to lure key developers and make a public or private fork of the project. This is a cheaper option, which also provides the necessary control over software assets.

The controversy between Amazon and Elastic is the result of a clash of models. To the credit of Banon and Elastic, Elasticsearch software has proven to be extremely popular, including through a licensing license.

However, permissions allow the system to be used by cloud providers such as Amazon. In order not to miss a profit and satisfy the needs of their customers, cloud providers will certainly offer native services for Elasticsearch and similar projects that are popular and well-known.

• Licensing is unrealistic. Despite the opinions of some investors, in fact, adding commercial terms to previously free software will never force the largest cloud service providers to subscribe. No company operating on such a scale will want to provide a major service — be it product development or pricing — to a third party that they do not control.
  
• Acquisition is another option to meet demand, but it does not scale well. Even wealthy cloud providers do not want to pay extra money for buying each new service in their portfolio, especially when there is a cheaper and simpler alternative - but here it is.
  
• In open source communities, fork has historically been viewed as a toxic option, but from a PR perspective, it becomes more acceptable if a commercial open source provider jeopardizes its own status by adopting tactics and methods that are contrary to the norms of the open source community. In this case, even large third parties may try to take a higher moral position while serving their own interests.

Faced with these options, the fork appears to be the logical response of the cloud service to the emergence of adverse licensing conditions. That's why Amazon’s decision was expected and inevitable. And therefore, it is difficult to identify the culprit in this situation. In principle, both sides acted logically - as one would expect, given their prospects, capabilities and legal rights.

It is likely that Amazon will be the first, but not the last cloud provider to do so. Others will also try to reconcile consumer demand with the absence of legal restrictions on the creation of their projects, such as the “open distribution for Elasticsearch”. They will probably inevitably conclude that this is beneficial. Apparently, commercial open source vendors will also inevitably conclude that the cloud is so big a threat that open source borders need to be expanded.

In fact, the only real question is whether the open source developers will draw a conclusion from the current situation with Elastic, which now competes with Amazon not only in products, but also in open source. Do they understand that the benefits of some controversial licensing approaches simply do not justify the costs.

However, maintaining the status quo is more likely. The incentives and motives of both parties are clear, understandable, and logical in the context of their respective models. Models that will always contradict each other internally, even if they are inextricably linked.

A hundred years ago, leaders of dozens of countries decided to enter into conflict. They knew that the conflict would cost them dearly, would be terribly destructive, and hardly anyone would emerge victorious from it. They did this because they did not see another way.

The technology industry does not seem to see either.

Note: Amazon and Elastic are RedMonk customers, as are Google and Microsoft. Expedia and Netflix are not RedMonk customers.