I attended the first Huawei Connect back in 2016. Since then, Huawei’s enterprise business has come a very long way. This year, neXt Curve was invited to attend the 6th annual conference which was held virtually from September 23rd to the 28th. This year, it is all about cloud native. The growth of Huawei’s cloud business is impressive when you consider that it was only four years ago that Huawei launched their Cloud Business Unit (BU). The BU’s achievements in advancing cloud technologies and services toward the edge is also nothing to sneeze at.
Back in 2016, I had an inkling of Huawei’s interest in getting into the cloud computing business. “Cloud” was an IT service delivery model that the company could not ignore especially as it looked to expand its enterprise IT business. At the time, the Enterprise BG (Business Group) was largely hardware centric selling network equipment and servers to enterprise customers. Getting into a cloud-based IT services business was a huge leap for the company. When the Cloud BU was launched as a subsidiary of the Enterprise BU in 2017, it was truly modest.
Fast forward to 2021, Huawei’s cloud strategy has permeated the DNA of the company. Over the last two years, it has become the platform for its AI endeavors and a fast-growing portfolio of AI services, tools and assets that constitute a core pillar of Huawei’s enterprise aspirations and strategy. The cloud is now Huawei’s modus operandi for consumer, enterprise IT, and network service delivery.
Huawei has made impressive strides in advancing their cloud agenda. Eric Xu, Rotating Chairman of Huawei, claimed that “Huawei Cloud has brought together more than 2.3 million developers, 14,000 consulting partners, and 6,000 technology partners, and also made more than 4,500 services available in the Huawei Cloud Marketplace.”
Huawei Cloud is currently the world’s fastest growing public cloud service provider and is ranked second in China and fifth in the world according to Gartner. Five years ago, Huawei’s cloud business was not even a blimp on the radar.
This rapid technological progress and cloud business expansion reminds me of the astonishing growth of Huawei’s Consumer BG which in a matter of a few short years put Huawei at the top of the smartphone market prior to the pandemic and souring of US-China relations.
So, is Huawei becoming a cloud business? If so, where is Huawei taking cloud? Well, here are the six takeaways that I tuned into at this year’s Huawei Connect.
- Ubiquitous & Open Intelligence
- Cloud to Cloud Native & Beyond
- Multi-Cloud, Multi-Edge Cloud Stack
- Everything as a Service
- Cloud + Cloud
- Cloud Technology Leadership through Open Source
Ubiquitous & Open Intelligence
For Huawei, artificial intelligence (AI) and the cloud go hand in hand. Not only is AI essential in realizing scalable, elastic digital infrastructure, it is an increasingly important modality of computing for enterprises and consumers. AI is also enabling a new breed of digital applications and services. Huawei leadership is clearly keen on providing the technology and infrastructure foundation for the new generation of AI-infused computing.
Huawei has made a great deal of progress building out their AI portfolio which was launched in 2018 and dubbed ModelArts. ModelArts is Huawei’s suite of AI system development tools, frameworks, and model assets that foster MLOps to simplify and accelerate AI application development. It is tailored to support the different levels of “AI maturity” and capabilities of an organization which Huawei defines as “experimental”, “quick win”, and “systemic”.
Much like AWS’s full AI stack and pipeline offering, Huawei is not just providing infrastructure, software, or tools. They have developed in short order what appears to be a highly integrated silicon-to-container/function platform optimized to make AI accessible, rapidly deployable, scalable, and energy efficient. This integrated stack approach has allowed Huawei to rapidly offer optimized development accelerators and assets such as Pangu large models which are massive pre-trained ML models that can be leveraged to rapidly build advanced AI applications.
ModelArts has also embraced the concept of no-code/low-code development for AI. Huawei clearly recognizes that a key barrier to AI adoption and benefit realization is the current paucity of AI developers and the steep learning curve that AI engineers face. To expand the AI developer community, Huawei seems to believe that the simplification of AI development will be essential in getting more developers in the learning cycle and organizations progressing beyond the “experimental” maturity phase.
Another principle that Huawei has embraced as part of their AI strategy has been making their technologies open. Huawei has introduced MindSpore AI computing framework in August of 2019 and made it open source in March of 2020 on Gitee and GitHub. According to Huawei, MindSpore has been downloaded over 600K times with 100 universities providing courses in their AI curricula. I suspect that going open source is an approach that Huawei believes will help accelerate the growth of their developer community and catalyze the global expansion of mindshare and market share for their AI technologies.
Cloud to Cloud Native & Beyond
This year, Huawei announced Huawei Cloud Native 2.0 services. A key element of this new service concept is Huawei Cloud UCS (Ubiquitous Cloud Native Services) which aims to take “cloud native” computing (or container-based computing) to the next level – distributed cloud native computing.
What is Huawei Cloud Native 2.0? In essence, it is multi-cloud container and serverless computing from central cloud to edge clouds that will provide “consistent user experiences that are not constrained by regions and traffic restrictions.” It also represents the transition from what Huawei calls “On Cloud” (deploying and running services on the cloud) to “In Cloud” (developing service on the cloud using a cloud-native architecture).
Huawei Cloud 2.0 moves the Huawei’s cloud vision sharply toward serverless computing. neXt Curve cited the emergence of serverless edge computing in our research for Ofcom’s long-term technology futures program more than two years ago. Our hypothesis was that serverless edge computing would be a transformative technology that would take edge computing and the prevailing concepts of MEC (Multi-Access Cloud Computing) concepts beyond container-based applications and deployments.
Why is this important? Serverless computing, especially in the context of edge cloud computing, opens a new frontier of innovation and new possibilities for designing and deploying distributed applications or functions. From the many sessions on serverless computing that I attended, it is obvious that Huawei is looking at serverless computing software and hardware technologies outside of the data center and from the perspective of edge computing for enterprise AND carrier networks. Huawei is uniquely positioned to drive serverless out across the edge with the potential of shaping the future of network software architectures and edge computing models.
One of the key benefits of serverless computing is deployment speed which is in milliseconds. This is in contrast with containers which require minutes to deploy initially. In the context of mobile edge computing, the near-real time placement and spinning up of functions opens new possibilities on how distributed applications can be design and deployed to take advantage of low latency 5G connectivity along the access network. Enterprise users will be able to use these capabilities to build novel edge computing applications along the emerging 5G edge clouds.
This year, Huawei showcased enhancements to their FunctionGraph offering, which is their serverless service platform that supports a wide range of application scenarios including real-time file processing, real-time stream processing, web and mobile backend, and AI functions such as OCR and lightweight AI inference. Each scenario leverages persistence services such as OBS (Object Storage Service) to enable stateful serverless functions. In this regard, Huawei is helping to change the perception of serverless computing being relegated to stateless functions.
Serverless computing is obviously a domain that Huawei continues to invest heavily in research and development as they push forward with their distributed cloud native strategy. It will be interesting to see how Huawei leverages advances in container and functions scheduling and orchestration to push serverless computing, in particular stateful serverless, out from the cloud data centers to the edge.
Multi-Cloud, Multi-Edge Cloud Stack
Over the last couple of years, Huawei’s cloud strategy has become very similar to IBM’s. If you can’t beat AWS, multi-cloud and hybrid cloud them. That is exactly what Huawei is doing and they are doing it with Huawei Cloud’s latest iteration featured at this year’s Huawei Connect event.
This year, ManageOne took the spotlight as the cornerstone of Huawei’s multi-cloud technology portfolio. It is Huawei’s cloud service management platform (CMP) that has evolved over the past couple of years into a multi-cloud brokerage platform for managing cloud services from across multiple cloud service providers as well as enterprise private clouds. This brokerage capability is increasingly important not only for enterprises who suffer cloud provider silos but for operators who are thinking about offering multi-cloud edge cloud computing services to their customers on top of their network.
One of the big challenges that enterprises face is vendor lock-in issues due to the lack of neutral open-source implementations of Kubernetes (K8). AWS, Azure, and other cloud service providers have their own flavor of K8. These vendor silos not only make it difficult for enterprises to manage microservices across vendor clouds, but they also make cross-cloud orchestration of containers and workloads challenging.
To address these issues, Huawei developed Karmada which is a multi-cloud middleware platform that enables multi-cloud container orchestration and management by mapping vendor-specific APIs to a vanilla (vendor-agnostic) API framework for managing K8 clusters across vendor clouds. Through this integration, Karmada can administer multi-cluster scheduling policies providing enterprises holistic control and management of containers across vendor clouds, or what Huawei calls multi-level cloud management.
Karmada also allows enterprises to reduce data gravity constraints associated with containers and apps being bound to data residing in vendor-locked clusters. According to Huawei, containers and data can be migrated across clouds in seconds fostering multi-cloud collaboration, multi-domain disaster recovery, and multi-architecture computing.
Huawei’s vision for multi-cloud container management and orchestration is an important direction toward realizing multi-cloud service management platforms for orchestrating containers and managing applications across the cloud continuum from data center to the edge. Not only will this capability enable enterprises to better manage their multi-cloud portfolios, but operators can also use these platforms to offer multi-cloud platform services (PaaS) to their enterprise customers on top of edge cloud infrastructure (IaaS) and network (NaaS) services that they host.
Everything as a Service
The foundation of Huawei’s cloud native strategy and vision is “Everything as a Service” (“XaaS”). Huawei’s interpretation assumes that all devices and infrastructure will be digital and will be cloud enabled. Huawei extends the XaaS concept to services such as Expertise as a Services for what they call “shared excellence”. This means creating tools, accelerators, and digital assets that capture and encapsulate organizational and domain expertise and offering it as a consumable digital service.
In essence, Huawei proposes to strip away the complexity of “digital transformation” to make it easier through low-code/no-code abstractions, composable AI and IT services, and many other constructs and methods that were mentioned across keynote and session presentations.
Huawei’s XaaS vision also includes the idea of simplifying the customer experience. Enterprise customers can manage their entire Huawei XaaS portfolio with one account, one interface, and one invoice. For retail verticals, Huawei is taking on a Shopify-like play by allowing its retail customers to build cyber-physical storefronts on Huawei’s XaaS stack and offering simplified interfaces and touchpoints to their customers for account management, ecommerce administration, and billing.
Operators could also benefit from Huawei’s XaaS offerings by building their own localized cloud services based on Huawei Cloud’s industry-specific software and platform services that Huawei and Huawei Cloud ecosystem partners and developer community are developing for enterprise customers. Huawei Cloud provides digital engines out of the box for payments, ads, browsing, maps, and search that are infused into their growing library of industry SaaS and PaaS offerings.
Huawei Cloud could also be a boon for operators that have largely struggled to expanding their portfolio of digital services to pursue new market opportunities in industrial and consumer sectors. Huawei’s one-stop-shop approach could accelerate an operator’s deployment of their own XaaS offerings. These services can also be optimized on top of an operator’s’ network by taking advantage of the converged IT/CT advancements in edge computing that Huawei is spearheading and implementing for its carrier customers.
Cloud + Cloud
A little over two years ago, when the US Commerce Department began a serial program of placing Huawei and other Chinese companies of concern on the BIS’s (Bureau of Industry and Security) Entity List and Google was prohibited from providing GMS (Google Mobile Services) installs on Huawei smartphones, Huawei went all in with their own mobile services platform called HMS (Huawei Mobile Services) and their own pan-device operating system dubbed Harmony OS.
In many ways, the US government’s sanctions on Huawei’s consumer business may have proven a boon for the company as a whole. The bans compelled Huawei to build out their own consumer device ecosystem with extreme urgency. In the span of a couple of years, Huawei claims that HMS’s monthly active user base has grown to over 800 million. Today, there are over 4 million developers in the HMS developer community, and HMS is now the 3rd largest mobile app ecosystem in the world with over 134,000 apps.
The rapid adoption of HMS and Harmony OS is not surprising. Google’s GMS applications and services are not essential in China, the world’s largest electronic device market. Moreover, Google’s GMS and search engine are banned in China and are restricted in 20 other markets around the world.
With the growing success of HMS and Harmony OS, Huawei announced “Cloud + Cloud” which is the company’s strategy to fuse Huawei Cloud B2B services and technologies with HMS Cloud B2C. The idea is that this union will enable “new business digitalization” for Huawei’s enterprise and carrier customers that capitalize on the expanding consumer base of Harmony OS and HMS users.
The basis of this Huawei Cloud business enablement initiative is formed by a simple arithmetic formula: HSM + Technology Innovation + Smart Operations + Global Capabilities + Localized Services. This simple math works out to what could become a very compelling and strategic B2B2C play for Huawei as well as for their struggling carrier customers who are looking to provide compelling localized digital business services to their own consumer and business customers.
Cloud + Cloud also includes what Huawei calls “digital enablement services” for content providers. This is a service capability that the company has been developing for several years to host OTT delivery of linear and streaming media. While Huawei’s traditional industry peers such as Ericsson and Nokia have either shed their media delivery service concerns or simply lack presence, Huawei is doubling down and going after this global market opportunity dominated by big cloud players and CDNs such as AWS and Akamai.
Cloud Technology Leadership through Open Source
Over the past few years, Huawei has made significant improvements in their software chops. It’s becoming very noticeable. Furthermore, the level of sophistication and forethought in R&D has reached what I consider a world class level. It’s a bit surprising given that many if not most of us still consider Huawei a network equipment manufacturer and gadget maker. They are not widely known for their burgeoning software prowess or their cloud business.
As mentioned earlier, Gartner ranks Huawei as one of the top 5 cloud leaders. Based on neXt Curve’s technology future research, Huawei is also a cloud technology pioneer. The company has been a major contributor in many areas that neXt Curve deems critical innovation areas such as edge cloud orchestration, serverless computing, hyperconverged data center architectures, interconnect, and more. Their technology leadership across many important domains was on display throughout the conference, though much may have gone unnoticed.
Projects such as Karmada, KubeEdge, and Volcano (which is a framework for enabling batch computing across K8 clusters that support massive parallel computing and highly scalable architectures for HPC and ML workloads), I believe, are strategic technologies that are addressing very big challenges in getting cloud native services and capabilities out and across the edge. One of the reasons that Huawei might land on these critical topics in their R&D efforts is the fact that they are looking deeply at converged IT/CT issues and technical gaps that will need to be addressed to improve cloud-native RAN and unleash the promises of 5G MEC and edge cloud computing.
Despite what seems to be a brewing tech cold war between the US and China, Huawei has ironically embraced openness. Maybe not Open RAN, but many of these technology initiatives mentioned above have been contributed as open-source projects. KubeEdge has been approved by the CNCF (Cloud Native Computing Foundation) as an incubation-level hosted project a year ago. Karmada was also accepted as CNCF’s first multi-cloud container orchestration project earlier this year.
Enterprise is supposed to be the big 5G opportunity, but it took the back seat at Huawei Connect 2021. It’s all about distributed cloud native. However, the edge cloud is a new frontier for everyone, the enterprise, the operator as well as the technology firms who are developing the edge cloud technologies. Much of the future of the edge cloud hinges on traditional operators and what neXt Curve calls the emerging UTB (Under-the-Bottom) players profiled in our research for Ofcom’s long-term technology futures study.
The UTB threat is not just for conventional-minded operators. The threat is very real for ICT players who continue to see the future of communications as data and the traditional notion of the network. In the past three Huawei Connect events, Huawei has demonstrated their emerging leadership in the converged IT/CT domain. It seems that adversity has led Huawei to realize competitive advantages from important technology bets that they have made in the past and have incubated in their R&D organization for quite some time.
Today, Huawei is one of very few companies in the world that has market leading ICT offerings and technologies that are integrated up and down the stack from silicon to app and from cloud to edge. These are technical and competitive differentiations that cannot be discounted because of the very compelling optimization and speed to market value propositions that they present to operators as well as enterprises and consumers.
I’m looking forward to Huawei Connect 2022 to see how far Huawei and its Huawei Cloud ecosystem takes cloud native and distributes it across the edge.