Harnessing the 11.11 Global Shopping Festival as a Launchpad for Alibaba’s Green Technology

13
Li Cheng, Chief Technology Officer, Alibaba Group

Each year on November 11, China hosts the world’s largest e-commerce bonanza, which is also one of the globe’s largest digital projects. Behind the scenes, an array of innovative technologies is put to the test. Alibaba Group, for example, has launched several new technologies to support its 11.11 Global Shopping Festival in recent years. These include a self-developed database (OceanBase) launched in 2014 for hosting transaction systems, the world’s largest hybrid-cloud architecture in 2015, and core transaction systems migrated to the cloud in 2019, to name just a few. This year, for the first time, the festival is being run entirely on the cloud.

Sustainability is another highlight. Using cloud-native technology, we have been able to reduce the computing resources required for every 10,000 transactions by 50%, compared to last year.

The latest technology breakthroughs in computing power and cloud services are expected to bring wider benefits to the environment and society at large. Cloud technology is an indispensable driver that significantly enhances computing efficiency and contributes to energy saving and sustainability in the digital economy.

In the context of the 11.11 Global Shopping Festival, Alibaba is driving the development of environmentally friendly cloud technology and pushing the product transaction chain toward a low-carbon economy, turning technology into an enabler for sustainable development. 

Every product, from manufacturing and trade to circulation and transportation, involves certain carbon emissions. The ongoing challenge is to reduce the volume and cost of carbon emissions throughout the industry chain. Taking the 11.11 Global Shopping Festival as an example, we have been investing our efforts in a number of areas.

Firstly, we made our data centers more eco-friendly. Reducing energy consumption in data centers has been Alibaba’s strategic priority for several years. We are using a much higher proportion of renewable energy resources, including wind power and liquid cooling in the data centers in Zhangbei, Hebei province in China and Renhe, Hangzhou where Alibaba is headquartered. By using clean energy in its data centers, Alibaba has reduced its carbon emissions by 26,000 tons during this year’s 11.11 Global Shopping Festival.

Secondly, we significantly increased the utilization rate of our data centers. In the past, utilization of data centers, especially the small ones, could be less than 10% of their capacity. Cloud operating models and technological architecture are facilitating more efficient deployment of computing resources, boosting the utilization rate for data centers. Holding an 11.11 Global Shopping Festival that ran entirely on the cloud with 100% cloud-native architecture, we saw a 20% efficiency boost in technology deployment, while CPU resource utilization increased by 30%.

Thirdly, we used more powerful chips to optimize computing performance and energy efficiency. This year, the scaled deployment of Hanguang 800, the first AI inference chip launched by Alibaba DAMO Academy (DAMO) in 2019, resulted in a 200% increase in algorithm performance for the search function on Taobao marketplace and a 58% reduction in energy costs.

Alibaba’s proprietary M6 model, debuted during this year’s festival, supported many AI-intensive operations. This world’s largest pre-trained AI model, with its high performance and low-carbon features, can generate design samples for human designers to reference and further improve. It is estimated that M6-powered design and production, along with the application of other eco-friendly materials, can reduce the carbon emissions that come from making a T-shirt by over 30%. 

Circulating and transporting goods releases CO2 as well, and this too can be addressed by technological innovation. For example, precise sales predictions can reduce pressure on supply chains and shorten inventory distribution distances; AI algorithms can prepare the right-sized packaging for parcels; and smart routing optimizes transportation routes for delivering the goods. Pairing this with the use of eco-friendly packaging material and smart order-consolidation algorithms, which combine multiple parcels to the same address into a larger package, can contribute to further reducing carbon emissions from delivery.

Reducing carbon emissions requires contributions from all sectors, and we have been sharing our green technologies with customers and partners over the past year. Taking the energy sector as an example, we are developing algorithms to help them optimize energy efficiency. The weather-dependent new energy sector will need precise forecasts for sunlight, wind speed, etc. to maximize output. The Short-term AI Weather Forecasting Platform launched by DAMO can provide precise “nowcasting” in real time to help energy enterprises plan and trade their carbon credits.

Enterprises nowadays should be shifting their priority from business growth to sustainable development, and they should be evaluated by social values, not just business values. As a result, the demand for “green computing power” will continue its exponential growth. Traditional heavy industries, such as iron and steel, electricity and coal, need to transit to low-carbon operations, while consumers have higher expectations for green logistics and low-carbon commodities. All of this will open up new opportunities, together with new challenges, for cloud services providers.

By leveraging technology to help reduce its carbon emissions, we believe this year’s 11.11 Global Shopping Festival was a launchpad for Alibaba’s broader green technology plans. We will continue this journey towards sustainability together with our partners, merchants, businesses, and consumers.