Catch the replay of the Apsara Conference 2020 at this link!
At the sub-forum on Smart Entertainment Industry Practices at the 2020 Apsara Conference, Saite, Senior Technical Expert of Alibaba Group, did a presentation. He talked about the technical exploration of Damaiwang in the entertainment industry with "one ticket technology".
The following content is a summary of his presentation.
Damaiwang provides various services for many scenarios, including concerts, venues, theaters, sporting events, large-scale exhibitions, and e-sports competitions. Many of them are characterized by high concurrency and large traffic.
Damaiwang's core business is closely related to "one ticket technology". Usually, the business procedure starts with the event organizer's commission of ticket production. Daimaiwang draws seat maps, sets ticket prices, and sells tickets through self-sales and distribution channels. Next, tickets will be delivered to customers. At last, customers come to watch the show with tickets. The procedure shows that the overall service chain is a relatively long closed loop from online to offline. Therefore, there are technical challenges in handling "tickets" in each stage.
Let's see the analysis of Damaiwang's business from the perspectives of people, goods, and places. Firstly, from the perspective of people, consumers are core roles that require a stable system supporting high-performance ticket-selling. Secondly, from the perspective of goods, content providers and distributors are core roles that require high efficiency and a good experience of seat selection and ticket production. Therefore, Damaiwang needs digitalization in seat management to improve customer experience. Finally, from the perspective of places, venue providers and regulators are core roles that require admission security and high deployment efficiency. Therefore, Damaiwang needs to provide on-site intelligent management. Based on the analysis above, the technical solution for the ticketing platform of Damaiwang should strongly support three core capabilities: high-performance ticket-selling, digitalization in seat selection, and on-site intelligent management. Now, I will analyze these capabilities of Damaiwang's ticketing platform in detail.
Most of the projects that Damaiwang undertakes are very popular events, such as large-scale concerts, sporting events, global e-sports competitions, and large exhibition activities. The number of tickets sold for some popular events can reach numbers similar to the number of products and transactions during Alibaba's Double 11 Global Shopping Festival. The large-scale projects of Damaiwang are characterized by big quantity, high risk, and low fault tolerance. To deal with these, Damaiwang pays attention to two core capabilities, seat selection capability, and inventory capability.
In terms of the seat selection capability, three issues need to be solved, namely, high concurrency, high data volume, and high collision rate.
For high concurrency, by using the concurrent-to-serial method, a snapshot is generated for each second-level timeslice and stored in the cache. Then, with "second-level timestamp", the corresponding request is mapped to the corresponding cache of timeslice snapshot. By doing so, the concurrency level can be reduced and the system availability can be ensured while satisfying the seat consistency. The process of snapshot generation and storage of timeslices under the second level can also be achieved through flexible configuration according to the parameters.
For high data volume, Alibaba Cloud adopts two-stage compression to reduce the data volume. In the first-stage compression, Alibaba Cloud uses its exclusive algorithm to compress the seat-related data based on actual needs. Data, like the seat ID, can be compressed into offsets, which will generate many duplicate strings, such as "012". Therefore, the second-stage compression is used. Alibaba Cloud uses universal compression algorithms, such as .zip and .7z, to compress these duplicate strings with a good compression ratio. After many tests on the client side, the decompression time is nearly negligible. A high compression ratio indicates that bandwidth consumption can be greatly reduced, and the user experience can be improved.
For high collision rate, it will occur when many users choose the same seats when booking tickets. The collision rate will be very high without corresponding selection restrictions. In extreme scenarios, the collision rate in the first minute of ticket booking will reach over 60%. Due to a large number of collisions, users have to return to the previous seat selection page to re-select, which brings great pressure on the system when the operation is repeated. To solve this problem, Alibaba Cloud has introduced the module of seat pre-locking. Generally, seat-locking is operated based on the database. This type of operation consumes lots of system resources and has a low concurrent performance. Therefore, Alibaba Cloud does not use this method.
However, the seat pre-locking is operated based on cache, which consumes a few system resources and has a high concurrent performance. In addition, DDoS attacks in seat pre-locking, such as occupying seats through malicious web crawlers, also need to be prevented. To solve this problem, Alibaba Cloud introduced risk control modules for identifying risks on the user side to prevent this kind of attack. The most important thing in the seat pre-locking is the setting of locking time. If the locking time is too long, the weakness of it will be easily amplified by machines used for occupying seats. As a result, tickets cannot be sold in a large quantity. If the locking time is too short, however, users cannot pay for their tickets on time. To solve this problem, Alibaba Cloud adopts the method of combining real-time and historical data to decide the locking time. The result shows that the collision rate in the first few minutes is reduced to less than 5%.
In terms of inventory capability, Alibaba Cloud has improved the performance and guaranteed stability in the following aspects.
First, the core procedure and non-core procedure are decoupled. Alibaba Cloud separates the inventory reading and the inventory deduction and degrades the weak dependency degree of the core procedure. By doing so, Alibaba Cloud generally ensures the stability of the core procedure of inventory operation from the architecture level.
Second, for the hotspot of inventory operations, a discrete algorithm for the hotspot is adopted. Based on business data, relevant inventory records are distributed to multiple tables of multiple databases, reducing the pressure of row-level locking in tables with single databases. In addition, Alibaba Cloud improves the performance of row-level record locking in a single row by introducing the hotspot patches used by Alibaba Group for Double 11. Moreover, the additional "concurrent deduction requests" is degraded to "serial deduction requests" when the inventory deduction of a database reaches a certain concurrent threshold. Thus, the high stability of the database can be ensured.
Third, it is necessary to ensure the quantity of inventory is accurate in real-time. Seats are different from clothing. Clothing is plentiful while seats are scarce. Once the number of event seats in an event is lower than what's showed on online platforms, you cannot solve the problem by simply adding more seats. Therefore, Alibaba Cloud ensures the consistency of inventory operations and prevents any inconsistent quantity by splitting large transactions and checking accounts in real-time.
"Seats" are entities in the core position of the entertainment industry. The production process of "tickets" includes seat maps drawing and box office planning. The ticket distribution process covers seat reservations and seat selection. The ticket fulfillment stage involves manual seat selection for supplement and machine-based seat selection. The ticket verification stage contains monitoring on attendance rate. All of these stages mentioned above are closely related to "seat". Therefore, Alibaba Cloud will pay special attention to processing efficiency and user experience.
With seat production and sales, by using an online editor based on SVG format, Alibaba Cloud can quickly record the location of physical seats on a seat map in SVG format. Drawing efficiency can be improved. With the SVG Codec module, the standard CSS specification is supported to ensure consistent rendering in different devices. By combining the VR seat map and the 2D seat map, Alibaba Cloud integrates the VR production, post-processing, mapping binding in the background into one processing flow. Thus, fast scale production of the venue's VR seat map is supported. In addition, by checking the SVG seat map and VR seat map, users can perceive the real perspective from their seat position for a greater experience.
In terms of seat selection and box office planning, many problems need to be solved, such as inefficient and long-time participation of employees. The single planning of large venues takes at least half a day, and the manual matching in large projects needs dozens of people each day. The work heavily relies on the working experience of employees. To solve these problems, Alibaba Cloud scores digitalized seats according to their basic attributes, such as position, angle, distance, and area. Based on certain rules and scoring algorithms, Alibaba Cloud gives general scores for every seat, row, and grandstand after calculation. By combining machine-based seat selection with box office planning, Alibaba Cloud determines the quality of a seat. In this case, machines can be used to replace manual work to improve business efficiency. By integrating manual processes into the system, Alibaba Cloud can retain years of manual practices for continuous accumulation and improvement.
What is the on-site management of Damaiwang? It focuses on the admission of customers and the prevention of gathering. On-site management has high requirements for stability. Through summarizing, four aspects need consideration: the complex on-site environment, gathering events, high stability requirements, and being perceivable during the whole procedure. To address these problems, Alibaba Cloud developed an efficient networking solution that supports various devices to deal with different ways of admission. Alibaba Cloud also uses on-site disaster recovery and visualization technology to analyze and perceive the on-site situation.
The traditional solution of efficient networking in large venues is "Access Point (AP) relay". AP relay can be realized by setting several interconnected APs around the venues to relay network signals. The disadvantages are the low efficiency of on-site manual configuration and low network stability. Once an AP stops working, the entire network will be unavailable. In the new solution, Alibaba Cloud developed the intelligent networking configuration of Maixiaozhi. Based on the LoRa protocol, Alibaba Cloud provided the local gateway service, enabling the automatic networking of gateways without excessive configuration. In addition, the local and cloud network channels support dual-channel mutual backup of 4G and NB-IoT. The new solution greatly improves on-site configuration efficiency and overall network stability and ensures high consistency between on-site local data and cloud data.
For networking devices, the transformation cost of adding original devices to the IoT module is relatively high. Therefore, Alibaba Cloud selected the independent gateway method. Alibaba Cloud used decentralization to ensure the independence of gateways. Thus, the failure of one gateway will not affect other gateways. In terms of 4G, Alibaba Cloud carried out network QCI optimization to ensure the stability of the 4G network of networking gateways even if the 4G service on the user side is unavailable. To ensure high consistency, NB-IoT is used to connect other data procedures, so network connectivity can be ensured by switching to the other data procedure when 4G service is unavailable. Multiple on-site devices are connected to the IoT platform for device management, so Alibaba Cloud can easily know the statuses of devices and perform remote device configuration. For the device SDK, Alibaba Cloud adopted software-hardware separation architecture. This architecture allows independent automation tests and is conducive to adapting new hardware.
For on-site disaster recovery, it is necessary to ensure that the whole admission system can still operate normally when the network is disconnected or powered off. Therefore, Alibaba Cloud adopted an integrated cloud-edge-device design. The same set of integrated SDKs for verification is used at the level of cloud, edge/local, and device. This ensures consistency in business rules and improves the efficiency of development. When the network is disconnected, a local system solution supported by device and edge will be used to support admission. When the power is turned off, the PDA device can independently support the admission verification. In addition, when the network connection is recovered, the related device data and edge data can be automatically synchronized with the cloud to ensure data consistency. The computing power of edge can also be used to support the computing in complex scenarios, such as face detection and crowd recognition. The relevant real-time data will be used for statistical analysis and transmitted to on-site screens, so the on-site staff can perceive the admission status in real-time.
An on-site IoT solution consists of three layers: device, edge, and cloud. By using integrated cloud-edge-device SDKs, the three layers are coordinated to ensure the consistency of rules and data. For device management, all on-site devices can be monitored and controlled remotely by connecting to the IoT platform. Based on the LoRa protocol, networking gateways implement local networking to improve configuration efficiency and network stability. In addition, networking gateways ensure the stability of 4G service under extreme conditions by optimizing the resource scheduling of the 4G network. Moreover, by connecting to the NB-IoT procedure, the dual-channel mutual backup of the final network resources can be ensured.
Damaiwang wishes to build a platform in the ticketing industry, which is called ticketingPaaS (tPaaS). There are three layers for tPaaS. The bottom layer is a cloud-based infrastructure. The middle layer is the middle platform capabilities of Damaiwang in the ticketing industry. The upper layer is the SaaS standard solution, which also supports customized development and output of individual technical capabilities. For tPaaS, Alibaba Cloud wishes to support deep customization of large projects by extending SPI and OpenAPI. Thus, the system capabilities can be continuously expanded and improved. By doing so, it is also convenient for Alibaba Cloud to attract more industrial eco-partners to better serve key customers.
What is the impact of new technologies on the ticketing industry?
First, blockchain technology can play an important role in making ticketing trustworthy and traceable. For example, it can prevent customers from purchasing fake tickets and solve the problem of buying tickets in large quantities. In addition, blockchain technology is conducive to the distribution of original rights and income in the copyright and IP derivatives.
Second, 5G technology brings brand-new online and offline interactive experiences to the industry, including online virtual scenes and offline interactions in concerts. This is especially true for offline interactions. Most offline interactions rely on costly bracelets based on IoT. With 5G technology, due to its high connection and low latency, cellphones can be used as IoT devices, improving the future possible forms of on-site interaction.
Alibaba Clouder - July 23, 2021
Alibaba Clouder - January 19, 2021
Alibaba Cloud Community - March 6, 2023
Alibaba Cloud Community - July 25, 2022
Alibaba Cloud Community - May 30, 2022
Alibaba Clouder - November 30, 2018
Provides secure and reliable communication between devices and the IoT Platform which allows you to manage a large number of devices on a single IoT Platform.Learn More
Link IoT Edge allows for the management of millions of edge nodes by extending the capabilities of the cloud, thus providing users with services at the nearest location.Learn More
A cloud solution for smart technology providers to quickly build stable, cost-efficient, and reliable ubiquitous platformsLearn More
An all-in-one service that provides elastic, stable, and widely distributed computing, network, and storage resources to help you deploy businesses on the edge nodes of Internet Service Providers (ISPs).Learn More
More Posts by Alibaba Clouder