Lost in the fog of TikTok was an important announcement: Huawei open sourced its homemade mobile operating system, HarmonyOS, now dubbed OpenHarmony. This announcement flew under the radar, but has far-reaching implications to not just the future of mobile technology, but also how that landscape will influence the geopolitical chess match between the U.S. and China.

But like many things that come out of China, it’s less a game of chess, but more a game of Go.

OpenHarmony: the Technology

Let’s first get a handle on the technology that underpins OpenHarmony. Because the project started, first in 2012 and intensified in 2019, as a strategic initiative to reduce Huawei’s reliance on the Android operating system due to U.S. sanctions, many mistakenly believe it is based on either Android or Linux (of which Android is based). That’s not true.

OpenHarmony is based on another open source operating system called FreeBSD. Interestingly, Apple’s macOS and iOS also leveraged FreeBSD indirectly from another operating system project called Darwin. So, on a bits and bytes level, Huawei’s OpenHarmony is more similar to iOS than Android.

This is a smart decision because if the whole point is to reduce reliance on Android, the most important thing to avoid is co-mingling code with either Android or Linux in case it triggers obscure licensing restrictions. And the last thing a development team wants is to have technology licensing lawyers checking every line of code the engineers write, especially when they need to develop quickly to avoid an existential crisis, which is what Huawei has right now.

Because the project is open sourced, there’s a lot one can do now to evaluate and verify OpenHarmony. In fact, I verified the FreeBSD lineage by rummaging through its codebase, which is all hosted on Gitee. For readers who had a chance to read my previous post “Can You ‘Nationalize’ Open Source?”, Gitee should sound familiar. It is a Git-based developer collaboration application that was recently anointed by the Ministry of Industry and Information Technology (MIIT) as the domestic “national champion” to drive open source growth in China. What I didn’t know when I wrote the previous post was that Huawei also became a strategic investor in Gitee in early September via its corporate venture arm, Habo.

But as every open source technologist can attest, no project ever gets traction without a long period of steadfast community-building and credibility-building. That’s where the OpenAtom Foundation comes in.

OpenAtom: the Foundation

The OpenAtom Foundation is China’s first non-profit organization of its kind geared towards fostering open source technologies, much in the same way as the Linux Foundation or the Apache Software Foundation. Huawei drove the founding of this foundation, and OpenHarmony is its anchor project.

But what is the point of a foundation anyway? As I’ve written in “COVID, Open Source, Industrial Policy”, a foundation’s involvement can help open source technologies in two meaningful ways: accelerate development and vendor neutrality. In a nutshell, vendor neutrality is important because it allows other large companies to contribute in the development of an open source technology without fearing vendor lock-in by another company, thus leads to faster development of that technology. An example would be Kuberentes, an open source container orchestration software that was first created by Google but is now the anchor project for the Cloud Native Computing Foundation (CNCF); Kubernetes’s fast growth would not be possible if it still resides within Google.

While a foundation’s involvement is by no means necessary -- and many open source projects have become popular without a foundation’s support -- it does help. And that’s what the OpenAtom Foundation is trying to deliver for China’s technology ecosystem. Its model and value proposition is similar to that of the Linux Foundation: basically delivering foundation (thus neutrality) as a service to open source projects, including legal, IP trademark management, licensing, community building, joint marketing, etc.

An open source foundation is successful when it builds an ecosystem of technologies around the anchor project with a coherent theme. The Linux Foundation, of course, built an ecosystem around Linux with many ancillary and adjacent technologies around the “open source operating system” theme.  The CNCF (a subsidiary foundation of the Linux Foundation) built an ecosystem of technologies around Kubernetes and the “cloud-native” theme.

While the OpenAtom Foundation is already hosting seven projects as part of its launch, with a lofty goal of fostering open source software, hardware, semiconductors, and content (I’m assuming documentation and technical education), the only theme seems to be that all projects were created by Chinese companies. And besides OpenHarmony, whose strategic value to Huawei is clear, the other technologies seem trivial to their original creators:

  • Xuperchain, a blockchain infrastructure project from Baidu
  • TKEStack (a container orchestration layer based on Kubernetes) and TencentOS (an energy-efficient IoT operating system) from Tencent
  • AliOS (a light-weight IoT operating system) from Alibaba
  • PIKA (a storage system based on the open source database, Redis) from Qihoo360
  • UBML (a Unified Business Modeling Language modeling system) from Netease

If I have to surmise a future theme that is technology-focused and not nationality-focused, it would be IoT because when OpenHarmony was first unveiled in 2019 (as HarmonyOS), it was an IoT-focused operating system, like TencentOS and AliOS. But since then, its scope has broadened to include support for smartphones, watches, and smart TVs.

One other curious element about OpenAtom is that only two projects (OpenHarmony, TencentOS) are hosted on the “national champion”, Gitee, while four others (AliOS, PIKA, Xuperchain, TKE) are on GitHub. The remaining one, UBML from Netease, requires a developer to fill out a form to apply for access, which is a very developer-unfriendly way to run an open source project.

Technology + Foundation = Developer Approval?

That’s the hope anyway. The nirvana of an open source technology, with or without a foundation, is to achieve widespread participation and buy-in among developers, who will both make use of the technology at scale and contribute to its development. And if an experienced foundation gets involved to leverage its best practices in open source management, it can increase the success rate by reducing much of the messiness and common mistakes that often plague young open source projects.

In the case of Huawei though, OpenHarmony is a young project, and OpenAtom is an even younger foundation.

The lazy and obvious conclusion here is to just dismiss all these efforts as a fool’s errand. But that’s not what you came to Interconnected for. The honest and nuanced conclusion is: it’s too early to tell and there are trends working both against and for Huawei’s open source strategy.

Factors Against

Huawei is infamous for its secretive ownership structure. Nobody knows exactly who or what owns Huawei. It’s a private Chinese LLC. It’s employee-owned, with 98.99% of company shares controlled by its employees via a “trade union committee”. Allegedly, this committee pays dues to more senior trade unions in an opaque bureaucracy that ultimately leads to the All-China Federation of Trade Unions, which is controlled by the Chinese Communist Party (thus all the controversy). That’s why I’ve advocated for Huawei to IPO in New York -- a bold act that would bring some desperately-needed credibility to the company.

Along the same vein, the OpenAtom Foundation also needs transparent, credible governance of its projects and itself. It currently boasts a 16-member Technical Oversight Committee (TOC), a typical governing element of an open source foundation, with Chinese technologists who have had years of experience working on projects in the Apache Software Foundation and Mozilla Foundation -- a good start. Their decision making process will have to be public and transparent to earn credibility from the wide developer community, both within and outside of China. As a reference, the CNCF TOC’s every governing deliberation is viewable and commentable on GitHub. Because OpenAtom is, after all, a China-registered entity, to what degree it can deliver pure transparency is questionable.

Lastly, OpenHarmony’s birthright as a Chinese creation makes building neutrality and credibility harder than just about any other birthright on the planet. This is an obvious yet important point that every Chinese company is struggling with right now. It is an element that every Chinese immigrant living abroad has been struggling with for much longer. For a young project, OpenHarmony does have reasonably good documentation in both English and Chinese -- an important first step that must be continued for the long haul. Maintaining a bilingual presence (much like this blog) requires lots of extra hard work -- work that an American-born, German-born, or French-born project does not have to do. None of us can pick where we are born, but we all have to deal with its uneven consequences. There is no point in pretending that doesn’t exist.

Factors For

It’s not all doom and gloom for Huawei; there are a couple of factors potentially working in its favor. For one, the U.S.’s own credibility and neutrality when regulating cross-border technology businesses is also deteriorating. The Trump administration’s wheeling and dealing of TikTok is nothing short of cronyism, so much so that it has been called out by none other than the WSJ editorial board. Although this doesn’t mean Huawei will have an opening to re-enter the U.S. market, other parts of the world may be more receptive to its technology. Regions like Southeast Asia, Latin America, the Middle East, and Africa are all credible possibilities. (See more in “Where Can the Chinese Internet Go?“ in the context of Alibaba and Tencent’s cloud expansion plans.)

From a product angle, Apple iOS’s current rift with Epic Games over the app store’s 30% revenue cut may present another opening. Regardless of which party is on the right side of history, Apple’s decision to suspend Epic’s access to its developer platform is already making other developers think twice about being part of the iOS ecosystem for the long term. Thus far, Apple has been able to lock-in the best developers within iOS because: 1. Its vertically integrated product approach delivers the best user experience; 2. People who buy Apple devices are more wealthy, so iOS users tend to have more money to spend, thus more money for developers to make. That may not last forever.

It’ll be years still before Huawei’s smartphones catch up to Apple’s, and sanctions around semiconductors against Huawei may further enlarge the gap. But if Apple’s goodwill among developers continues to fall, OpenHarmony could step in if its rollout into Huawei smartphones in 2021 goes well and design a more favorable incentive structure with developers. The economically rational developers will always look for an ecosystem with credible technology (open source), a big install base (Huawei is currently the #1 smartphone maker in the world), an audience that’s growing wealthy and spend money (the Chinese middle class), and friendlier revenue-split terms (TBD).

To be sure, the “factors against” are more immediate challenges, while the “factors for” will take a long time to play out. Does Huawei have the money to afford the time and patience?

A Multi-Century Game of Go

While the future is always hard to predict, especially when it comes to Huawei, the company does have a decent pile of cash to last through some very lean years, while executing this open source strategy. Huawei has about $53 billion USD in cash and short-term investments at the end of its fiscal year 2019.

To put this amount in perspective, here are some peer companies’ cash level at end of their respective fiscal year 2019:

None of this accounts for any government subsidies that may be injected into Huawei, which may never show up in its financial statements. Thus, Huawei should have the runway to execute its open source strategy to gain developer mindshare and technology market share without ever needing to make a dime from open source.

For anyone with just an amateur understanding of the game Go, you’d know that the pieces are all the same -- same power, same utility. It’s where the pieces are placed on the board that determines who will win the game. It’s also a game of relative advantage, not absolute dominance, that can take many turns to play out.

Huawei has clearly decided to place an important piece on the board that represents open source, with more pieces to come.

Huawei CEO Ren Zhengfei -- loathed as a Community Party puppet in the U.S. and revered as a generation-defining entrepreneur in China -- famously said earlier this year that Huawei’s operating system will take a long time to surpass either Android or iOS, but it won’t take more than 300 years. Except for Jack Ma (wants Alibaba to last 102 years and span three centuries) and Jeff Bezos (currently building the 10,000 year clock), not many tech CEOs, or any CEO, speak and act in a century-spanning timeline.

If we take Ren’s words both seriously and literally, then Huawei’s open source strategy will be something that will play out not in decades, but in centuries.

If you like what you've read, please SUBSCRIBE to the Interconnected email list. To read all previous posts, please check out the Archive section. New content will be delivered to your inbox (twice per week). Follow and interact with me on: Twitter, LinkedIn.


300年:华为的开源战略

在围绕TikTok的迷雾中,我们忽略了一条重要的新闻:华为将其自制的移动操作系统 HarmonyOS开源了,改名为 OpenHarmony。这一宣布不仅对移动科技的未来,而对中美地缘政治博弈的未来都有深远影响。

但和在中国发生的许多东西一样,这个战略不是一盘象棋,而是一盘围棋。

OpenHarmony的技术


首先让我们了解一下OpenHarmony的技术基础。次项目始于2012年,并在2019年加紧开发,它的战略目的是为了减少华为对Android操作系统依赖,尤其在美国制裁日益增加的环境下。许多人误认为该项目是基于Android,或是基于Linux(Android基于Linux),但其实并不是。

OpenHarmony是基于另一个开源操作系统FreeBSD。有意思的是,苹果的macOS和iOS也间接地用了另一个基于FreeBSD的操作系统,叫Darwin。因此,从纯技术角度来看,华为的OpenHarmony更像iOS,而不像Android。

这其实是个明智的选择,因为如果战略重点是减少对Android的依赖,那么最重要的是避免代码与Android或Linux混在一起,以防触发某些模糊的使用许可限制。一个开发团队最不喜欢的就是让科技授权律师检查工程师写的每一行代码,尤其是在需要快速开发迭代来避免生存危机时的状况下。华为目前的生存危机是显而易见的。

因为这个项目开源了,OpenHarmony的很多方面可以更客观的去评估和验证。我其实就是通过翻查它的代码来验证它与FreeBSD的“血统关系”。所有代码都托管在Gitee上。对有机会看了我之前的文章《能把开源“国有化”吗?》的读者, 对Gitee应该很熟悉。这是一个基于Git的开发者协作平台,最近被工信部评为“国家队”成员来推动中国开源的发展。我在写上一篇文章时还不知道的是,华为也在9月初通过其企业风投部门哈博成为了Gitee最新的战略投资者

但正如每个开源技术专家都可以验证的,没有长期稳定的社区建设和信誉建设,任何一个项目都不会什么影响力和吸引力。这就是OpenAtom基金会的用武之地。

OpenAtom基金会

OpenAtom 基金会是中国第一个致力于培育开源科技的非营利组织,与Linux基金会或Apache软件基金会非常相似。华为推动了这个基金会的创立,而OpenHarmony则是它的核心项目。

基金会到底有什么意义呢?正如我在《COVID,开源,工业政策》中所写到的,基金会的参与可以通过两种方式帮助开源项目:加速进度和厂家中立。简单的说,厂家中立性很重要,因为它允许其他大公司参与开发中立的开源技术,而不必担心被另一家公司锁定,从而加快该技术的发展迭代。一个例子就是Kuberentes,它是一个开源的容器编排软件,最初由Google创建,但现在是云原生计算基金会(CNCF)的核心项目;如果Kubernetes仍然驻留在Google内部,Kubernetes现在的快速发展是不可能达到的。

虽然基金会的参与不是必须的,也有许多开源项目在没有基金会支持的情况下发展蓬勃,但它确实有帮助。这也正是OpenAtom基金会试图为中国的科技生态系统提供的。它的构架和价值与Linux基金会相似:基本上将基金会(即中立性)作为一套服务提供给开源项目,包括法律、IP商标管理、许可、社区建设、联合营销等。

一个开源基金会的成功在于它能否围绕一个核心项目,一个科技主题来构建一个技术生态。Linux基金会就是围绕“开源操作系统”这个主题,用许多辅助和邻近技术构建了一个围绕Linux的技术生态。CNCF(也是Linux基金会的一个附属基金会)也建立了一个围绕Kubernetes和“云原生“这个主题的技术生态

虽然OpenAtom基金会已经托管了7个项目,作为其启动的基础,也有很远大的目标:培育开源软件、硬件、半导体和内容(我假设是文档和技术教育),但唯一的主题似乎是所有项目都是由中国公司创建的。而且除了OpenHarmony对华为的战略价值显而易见之外,其他项目对其原创公司来说似乎微不足道,并不重要:

  • 百度的区块链基础设施项目Xuperchain
  • 腾讯的TKEStack(基于Kubernetes的容器编排层)和TencentOS(一种节能的物联网操作系统)
  • 阿里的AliOS(轻量级物联网操作系统)
  • 奇虎360的PIKA(基于开源数据库Redis的存储系统)
  • 网易的统一业务建模语言建模系统UBML

如果我不得不预测未来以技术为中心,而不是以国籍为中心的主题会是什么的话,那就是物联网,因为当OpenHarmony在2019年首次发布时(当时叫HarmonyOS),它是一个物联网操作系统,就像TencentOS和AliOS一样。虽然从那以后,它的范围扩大到包括对智能手机、手表和智能电视的支持。

关于OpenAtom另一个奇妙的事情是,只有两个项目(OpenHarmony、TencentOS)在Gitee上托管,而另外四个项目(AliOS、PIKA、Xupercain、TKE)都在GitHub上。剩下的一个来自网易的UBML,它要求开发者填写一个表格来申请访问,可以说是对开发者最不友好的一种运营方式了。

技术 + 基金会 = 开发者的认可?

起码希望是这样。不管有没有基金会的参与,一个开源项目的最高点就是赢得广大开发者和程序员的大规模使用和贡献代码。如果有一个有经验的基金会来参与和提供开源管理中的最佳实践,那可以减少很多年轻开源项目的混乱和常见错误,从而提高成功率。

但就华为而言,OpenHarmony是一个年轻的项目,而OpenAtom则是个更年轻的基金会。

一个懒散而毫无创意的结论就是,对华为的这些努力不予理睬因为困难重重。但这不是您来看《互联》的目的。既诚实又有些细微价值的结论是:现在说还为时过早,而且有一些趋势既反对也支持华为的开源战略。

不利因素

华为有着自己独特而神秘的公司构架,众所周知。没有人确切地知道到底谁拥有和控制华为。华为是一家私有的有限公司,以员工持股著称,98.99%的公司股份由员工通过一个“工会委员会”控制。据称,该委员会会向更高级别的工会缴纳会费,最高一层是由中国共产党控制的中华全国总工会(因此引起了当前对公司的各种争议)。这就是为什么我曾经主张华为在纽约上市——这是一个大胆的举动,会给它步一些急需的公共信誉。

同样,OpenAtom基金会也需要对其项目和自身进行透明、可信的治理。它目前拥有一个16个成员的技术监督委员会(TOC),这也是一个开源基金会的典型管理体制,里面有许多中国科技界里参与过在包括与Apache软件基金会和Mozilla基金会项目中做开源工作的资深人士,这是个好的开端。他们的决策过程必须是公开和透明的,才能赢得中国国内外广大开发者群体的信任。作为参考,CNCF TOC的每一个大大小小的决策的过程都可以在GitHub上查看和评论。因为OpenAtom毕竟是一家在中国注册的组织,它能在多大程度上提供纯粹的透明度是值得怀疑的。

最后,OpenHarmony作为一个中国人创造的东西,从“出生那天”就在建立中立性和可信度上打了折,比出生在地球上其他任何一个地方都要困难。这是每个中国公司目前都在面对的一个挑战。这也是每一个生活在海外的中国移民一直在面对的长期挑战。对于一个年轻的项目来说,OpenHarmony已经有些不错的中英文双语文档,这是一个必须长期持续的重要一步。保持双语(就像这篇博客一样)需要很多额外的努力——这是一个在美国出生、在德国出生或在法国出生的项目不需要做的工作。我们谁也不能选择自己的出生地,但我们都必须面对其不均衡的后果。假装这种现象不存在是没有意义的。

有力因素

也并不是所有华为面对的都是挑战;有几个因素对华为还是有利的。首先,美国自身在监管跨国科技公司时的公信力和中立性也在降低。特朗普政府对TikTok整个交易的处理好比香港黑帮,以至于连《华尔街日报》的编辑部都忍不住要公开指责这个现象。当然这并不意味着华为将有机会重新进入美国市场,但世界其他地区可能会更愿意接受它的技术。东南亚、拉美、中东和非洲等地区都有可能。(可以看《中国互联网还能去哪里?》这篇文章,以阿里和腾讯云扩张计划的背景更深入的了解这个话题。)

从产品的角度来看,苹果iOS目前与Epic游戏在App Store的30%提成上的分歧可能会带来另一个契机。不管哪一方正确的,苹果决定暂停Epic对其开发者平台的使用权限,已经使其他开发者质疑长期在iOS生态里开发还是不是个明知的选择。到目前为止,苹果能够在iOS内锁定最优秀的开发者是因为:1. 它的垂直集成产品开发提供了最好的用户体验;2. 买苹果产品的人一般更有钱,因此iOS用户往往更能花钱,因此开发者喜欢为他们做产品来赚更多的钱。这种状况也许不会永远持续下去。

华为的智能手机要赶上苹果还需要数年的时间,而针对华为的半导体制裁会把距离拉的更大。但如果苹果在开发者群体中的声誉继续下降,如果OpenHarmony在2021年在华为智能手机启动使用进展顺利,并给开发者设计一个更有利的激励结构,OpenHarmony 可能会有发展空间。理性的开发者总是会寻找一个可靠的技术(开源),庞大的安装基数(华为目前是全球第一大智能手机制造商)、一个日益富有并爱花钱的受众(中国中产),以及更友好的收入提成安排(拭目以待)。

实话实说,这些“不利因素”都是眼前的巨大挑战,而“有力因素”则需要很长时间才能看出效果。华为是否有足够的资金来支付时间和耐心?

一盘跨世纪的围棋

虽然未来总是很难预测,尤其是对于华为而言,但公司确实有一笔可观的现金可以维持未来不景气的几年,同时继续执行这一开源战略。截至2019财年末,华为拥有约530亿美元的现金和短期投资

为了正确看待这一数额,以下是一些同行公司在其各自2019财年末的现金:

这里当然不包括华为可能从政府得到的任何补贴,这些补贴也可能永远不会出现在华为的财务报表中。因此,华为应该有机会落实其开源战略,在不需要拿它赚一分钱的情况下赢得开发者的”思维份额“以及技术的市场份额。

对于任何一个对围棋有粗浅了解的人,都会知道所有棋子都是一样的——同样的力量,同样的效用,谁赢谁输取决于棋子放在棋盘上的位置。这也是一个相对优势的游戏,需要多次轮番才能比出高低,而不是以绝对统治对手为目标的游戏。

华为显然决定在棋盘上代表开源的位置放下了重要一颗棋子,而且还会有更多的棋子跟进。

华为CEO任正非是个极有争议的人物,在美国被视为共产党的傀儡,在中国被尊为一代代表性的企业家。他在今年早些时候曾说过,华为的操作系统将需要很长时间才能超越Android或iOS,但不会超过300年。除了马云(希望阿里巴巴能持续102年,跨越三个世纪)和 Jeff Bezos(正在建造“万年钟”)以外,没有多少科技公司的CEO,或任何行业的CEO,会用跨世纪的时间线说话和行动。

如果我们既认真又直观的对待任老的这番话的话,那华为的开源战略将不会在几十年内,而是几个世纪内出结果。

如果您喜欢所读的内容,请用email订阅加入“互联”。要想读以前的文章,请查阅《互联档案》。每周两次,新的文章将会直接送达您的邮箱。请在TwitterLinkedIn上给个follow,与我交流互动!