Some weeks ago The Stack highlighted Apache EventMesh as the latest in our series on emerging open source software (OSS) projects in the Apache Incubator -- the project is heavily backed by China's WeBank.
China's companies are clearly taking the open source space seriously and with increased initiative, this week's featured project, Apache BlueMarlin, is also backed by a Chinese company -- Huawei subsidiary Futurewei.
The devs putting their efforts into the nascent Apache BlueMarlin are targeting a big prize: the multi-billion global adtech market -- controversially dominated by a small handful of major players running proprietary systems, including Google and Facebook in the US; Tencent and Baidu in China. They have their work cut out...
What is Apache BlueMarlin?
BlueMarlin officially went the open source route on October 8, 2019 and its R2.1 version was released to the public on January 15, 2020. It was voted into the Apache Incubator (the primary route to becoming a full-fledged project in the non-profit open source Apache Software Foundation) in 2020, but information on the project remains thin, with no website yet and a blank page on its GitHub repo dedicated to design documents.
Describing itself as a "critical web infrastructure for advertising based monetization", its team call it a "cloud platform that adds intelligence to a plain Ad System." Committers include developers affiliated with Futurewei Technologies, Intel, Alibaba, and Tencent, its GitHub reveals, although it has yet to publish any packages or releases onto the code repository that we could see.
Benefits of using Apache BlueMarlin
Digital advertising measurement and data collection now extend much beyond cookies. Big brands will need cloud-based architectures to effectively measure campaign performance as it helps in data mobility, agile functioning, improving scales and in building custom applications.
See also: What is Apache AGE? (#1 in our emerging open source series).
What the project does note is that Apache BlueMarlin aims to offer components like:
- Smart booking for inventory
- Performance forecasting across various regions, time zones and devices
- Advertising spending recommendation
- Optimise advertising-based monetization
- Competitor ad spend analysis and information on their products, prices, reviews and ratings
- Cross-device targeting of customers across mobile, desktop, tablets, smart devices, and wearables
- Identifying patterns in the advertising data and suggestions for changes to current campaigns
- Ascertaining effectiveness of various marketing activities within their media mix
- Acceleration of decision making in real-time from weeks to hours based on data inputs and its processing
Introducing Apache Sedona: Where ‘Big Data’ meets geospatial data.
The project participants' future plans include developing optimisation algorithms to allocate bookings in inventory, and deep learning models to suggest the best inventory based on pCTR, its developers suggest. With just a handful of pull requests each month and limited activity, it's hard to tell how much appetite Futurewei, as a primary sponsor, has for the project, but we will watch its progress with interest.
n.b. Those interested in building advertising platforms could do worse than to take a look at this refreshingly detailed write-up from Google Cloud.
.
