Knowledge graph technology has gradually become the outlet for AI, and graph databases have also stood on the forefront of the database field. As the world's first time series dynamic graph data warehouse (a new database storage architecture of time series + graph + data warehouse), AbutionGraph, Let's see what changes can be brought about in the field of telecommunications.
In various industries, the exponential growth of data has always been a sign of business in the 21st century. Organizations must find new ways to manage their large and growing data stores while transforming their most valuable assets (ie data) into actionable business insights. The arrival of 5G has further magnified the importance of data, especially in the telecommunications field. Mobile operators are facing fierce competition if they want to stand out. At the same time, the Internet of Things, smart devices, and new content platforms also make telecommunications companies face data shocks.
The arrival of 5G also means that with the proliferation of equipment, telecommunications are finding themselves in an increasingly challenging position, because 5G has become another catalyst for enterprises to achieve extreme data growth. At the same time, these organizations must deal with more and more use case scenarios, each of which has its own different needs, and at the same time faces technical obstacles to successfully access and analyze their data treasures.
Smart mobile operators realize the need to take advantage of the opportunities brought by 5G wireless technology, while creating new business models, optimizing network service quality, and launching new personalized products and services to their subscribers. The rapid growth of data (if properly managed) is the key to increasing competitive advantage and revenue. A recent study estimated that by 2035, 5G technology will be a driving force for $13.1 trillion in goods and services. Through the realization of critical communications, large-scale Internet of Things, and faster and better broadband, 5G will become a game changer. In order to promote business, every company needs to adapt to the 5G digital world of the Internet of Everything.
The huge data potential of 5G brings huge commercial value, making this potential a real obstacle to profitability: According to recent industry estimates, only a small part of the organization's data can be accessed and analyzed in the end. This means that the telecommunications industry is losing business opportunities that may change the rules of the game. This may be due to a variety of reasons: from insufficient budgets to constantly accumulating historical data in the database, to extremely long-running queries that cannot be executed (especially base station data, a large number of user access data), the value of these data itself It is extremely low and requires operators to fully perform data association and calculation analysis to effectively display their commercial value. So, what measures can mobile operators take to change this situation? How does telecommunications use explosive 5G data to gain a competitive advantage?
Generally speaking, the telecommunications industry follows the following basic steps when solving data problems:
1. Strengthen anti-fraud measures, locate marketing and add new services to reduce user loss;
2. Ensure that subscriber billing at each level reflects actual usage and performs profitability analysis;
3. Reduce costs by optimizing routing;
4. Provide analysis for adjacent markets such as content providers.
For most of the data that people and businesses use every day, telecommunications companies are both channels and links. If you want to use these data, you need to analyze the data. Telecommunications service providers must analyze real-time data and historical data at the same time, so that they can keep abreast of the situation when needed.
To solve many analytical challenges, telecommunications companies have tried a variety of techniques. For example, technologies such as in-memory database ES can provide real-time analysis for real-time data, big data technologies such as Hadoop can be extended to support a large number of historical data sets, and NoSQL databases (Hbase, etc.) can help enterprises expand to support data surges economically and efficiently.
However, the cost of in-memory databases is often too high and cannot meet the needs of telecommunications companies to analyze PB data. Big data cannot support real-time analysis, and NoSQL databases lack the characteristics of SQL databases, thus failing to provide rich analysis functions. Telecom companies have large-scale equipment. For data monitoring and analysis at the Internet of Things level, the above technologies obviously cannot meet these high-efficiency tasks. It is necessary to increase the technical direction to expand the business. This is the time series database (TSDB, etc.) and real-time data. The combined application of warehouses (Druid, Kylin, etc.) is reflected.
If these technical systems are fully utilized regardless of cost, a lot of data analysis can be completed in the telecommunications field, such as:
1. Provide a seamless experience for subscribers, including analysis to help optimize routers based on geographic location.
2. Quickly transfer relevant data between content providers, advertising providers and users' applications.
3. Support the real-time monitoring of the equipment by service personnel to troubleshoot problems, and allow viewing of historical modes to recommend the best plan or upgrade.
4. Ability to view routing data in real time and optimize network infrastructure, thereby reducing problems. At the same time, you can examine historical issues to determine whether a problem is an accident or a performance of the infrastructure reaching the design limit.
The fact is that telecom providers currently use many different technologies and platforms to perform these analyses. Since no traditional solution can meet the analysis needs of these companies, and no database can meet the multi-scenario business at the same time, most companies are New technologies are being explored to reduce the maintenance of the technical system, but this method is not sustainable in the long run. Telecom providers will face increasing hardware, space, power, and cooling costs. Cross-platform management and integrated analysis are also becoming more and more complex, which is more error-prone and costly.
Telecommunications companies must continue to innovate to achieve two opposing goals: one is to collect and analyze large amounts of real-time and historical data, and the other is to reduce the number of separate analysis platforms to control costs.
In order to turn data challenges into opportunities, most telecommunications providers are innovating through research and development analysis technologies. We use the technical solution of AbutionGraph to find the answer to the problem, and all of this will be significantly improved.
AbutionGraph is a graph data warehouse developed by the domestic Thutmose company. It is also the pioneer of the graph data warehouse. It unifies the three originally unrelated database technologies of data warehouse, time series database and graph database, which is a natural knowledge. Database technology, associated search engine, real-time data analysis and IoT data monitoring system. The emergence of such a technology is exactly what telecom providers need, and the high cost of cross-platform management and integrated analysis can be easily solved.
We must have some understanding of the data warehouse. Let's take a look at the conveniences that the "graphs" in the data warehouse can bring. Graph is knowledge graph, not a new technology, it belongs to AI category (neural network is also a large knowledge graph). It has existed as early as when Google was founded. It is used for web search engine and it is more of memory computing. The "graph" in the AbutionGraph graph data warehouse, we represent the storage of knowledge graph data, coupled with the characteristics of the time series database, that is, the time series dynamic knowledge graph technology. The research of time series knowledge graph has been included by the Ministry of Science and Technology in the next ten years-Technology Innovating the 2030 development outline, it can be seen from the national strategy that this is a technology that can change the intelligence of the world, and it is obvious that AbutionGraph has already taken the lead in the forefront of the world.
In short, the graph technology has some recognized advantages over traditional databases:
1) The fastest access speed
: Millisecond-level query response can be achieved in the graph database of trillion-level edges.
2) The strongest expressiveness
: Describe the various entities that exist in the real world and the relationships between them.
3) Multi-layer relation retrieval is the fastest
: Real-time (sub-second) link traversal (>3 levels), such as knowing whether my friends and Xiao Li’s friends know each other.
4) Support graph algorithm
: User group grading, such as: group precision marketing for users with the same surfing habits; quickly calculate the 10 people with the most similar behavior characteristics to Xiao Li.
5) Visual traceability
：Natural explanatory structure, can understand user behavior habits through graph visualization, and explain the results of algorithm models.
6) Explore new business directions
: Based on the graph structure, it is more convenient to discover new user anomalies, understand the overall trend of the data, and support business decision-making.
As shown in the above map model, we have built a call and person association map with users as the core, which can manage all the IoT smart devices in the home (for easy understanding, it can correspond to the complex telecommunication equipment network). At the same time, we It can also detect the call behavior habits of all telecommunications users, which can help content providers, advertising providers, and users' applications quickly transfer relevant data. The call relationship is the embodiment of AbutionGraph's time series. The use of time series of multi-dimensional user behavior modeling can further help combat cybercrimes, analyze the behavioral characteristics of online fraud gangs, and reduce corporate and personal property losses.
The emergence of AbutionGraph can bring more business exploration possibilities to the telecommunications industry with large-scale data. If a real-time data volume of nearly 100 million per second is built into a telecommunications profile network, it is difficult to control using traditional technology. Even if the data is already stored in the database, then data analysis will be a big challenge. If you can't check it, dare not check it, and the correlation analysis may collapse. This is the most true feedback from our customers. The large number and rich types of data are a major asset of telecommunications companies, and it is also a double-edged sword. While improving their own user services, convenient communication tools also bring huge benefits to the black industry. Various mobile phones/accounts, The content of the short message (QQ/WX number, URL) is scammed in endlessly. We use AbutionGraph to help telecommunications companies effectively combat cybercrimes. Its high timeliness can prevent risks before they happen. Large-scale real-time data combined with historical data analysis becomes easy.
In addition to common call records, telecommunications also has data such as text messages and Internet logs. Knowledge graphs have natural knowledge fusion. In AbutionGraph, each graph can be stored independently, and multiple graphs can be directly associated with queries when needed. This is knowledge fusion. Characteristics, we can simply and clearly perform fusion analysis on these multi-source data through knowledge graph modeling, and discover the associations among them.
It can be seen that AbutionGraph's solution is market demand to promote product innovation. It solves the weaknesses of previous big data technology, data association and real-time multi-dimensional/multi-label time series analysis. The telecommunications company is one of the first companies to deploy technologies such as memory databases and Hadoop. These technologies can perform real-time and historical data analysis, and can also be expanded under the constraints of hardware and management complexity. The new technology AbutionGraph can fully and The combination of these old technologies, such as reusing the Hadoop system to persist data, providing efficient data analysis and data governance solutions, and combining AbutionGraph with the seamless interface of Flink/Spark, can create a high performance without overthrowing the old big data platform A big data platform that can meet business update iterations.
The use of AbutionGraph and knowledge modeling of telecom data for business enables telecom companies to overcome three key challenges:
1) Maintain customer loyalty
Use the relevance of knowledge to find ways to expand core services and cross channels to maintain customer loyalty.
2) Reliable data service
Use the knowledge graph to analyze network data to detect and prevent network problems to provide high-quality data and call services.
3) Prepare for the impact of the Internet of Things
Traditional infrastructure and data warehouses cannot keep up with the scale of the Internet of Things, but only a time-series dynamic graph data warehouse can be satisfied.
Finally, industry customers indicate that at present, operators can only access and analyze a small part of their data. Because they cannot support exponential data growth and rely on traditional database systems, the telecommunications industry has left a deep accumulation of data. Insight is enough to change the rules of the game. Therefore, data is one of the most critical components of the telecommunications competitive arsenal, and the arrival of 5G will magnify its importance. Using knowledge graph + real-time data warehouse can quickly analyze a large amount of telecommunications data and store unlimited business opportunities. AbutionGraph has passed the real-time analysis and testing of large-scale telecom data, used to overcome technical challenges and successfully store its massive data in the enterprise, and discover key risk and value intelligence, effectively improving the value of operators' data utilization. ????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????? .