Lead:
Just after midnight, the express has arrived. Although there are traces of marketing behind this, it must be said that Double Eleven is no longer the previous Double Eleven. What is supporting Double Eleven becoming faster? There is no doubt that technology, especially real-time analysis technology represented by memory computing, is making business activities traceable and evidence-based.
The annual Double Eleven Shopping Festival is not only a shopping carnival day for consumers, but also a severe battle test for e-commerce, logistics, supply chain, and brand owners: how much sales can be expected on Double Eleven this year, and how to achieve sales accuracy Forecast, so that there is no shortage or accumulation? How to optimize dynamic pricing to maximize revenue and profit? How to provide consumers with precision marketing and achieve personalized services? How to optimize the supply chain and make production, warehousing, and replenishment calmly...
Compared with previous years, in the last one or two years, different companies have begun to become more distinct when fighting the "Double Eleven". Some companies are still as busy as ants on a hot pot, but they have no clue or tactics. The Double Eleven has hurriedly ended amidst the turmoil, but will repeat itself again in the coming year; while other companies are busy, but quite a bit. Calm and self-confidence, all arrangements are in order, orders flew in between talking and laughing, and even the subsequent double twelve and year-end promotions are clear.
The root of this is: companies that embrace data analysis capabilities are beginning to benefit from it. These companies try to build powerful data analysis models in marketing, operations, supply chain, production and other links, and use artificial intelligence technologies such as machine learning and memory computing technologies to continuously train, learn and analyze based on data, so as to be in the rapidly changing market The environment provides accurate support for real-time decision-making in business links such as marketing, operations, and supply chain.
In fact, the introduction of AI and memory computing technology to bring changes to enterprise data analysis is considered to be the general trend. Garnter predicts that in the next three years, the popularization of data analysis knowledge and technology will accelerate, allowing employees of major companies and organizations to use advanced data analysis functions to obtain valuable insights from data. From the company's overall strategy to specific business scenarios, companies that fully embrace data analysis capabilities will take the lead in future market competition.
1
AI and memory computing revolutionize data analysis
In the traditional data warehouse era in the past, data analysis usually had a single data type, poor real-time analysis, and support for local businesses. This data analysis model was difficult to truly realize "data-driven business". Now, data analysis is not what it used to be. General public employees can make various decisions based on data insights in different daily scenarios, thanks to the rapid development of data analysis related technologies, such as machine learning, natural language processing, etc. AI technology is integrated into data analysis.
One of the representative technologies is enhanced data analysis. Garnter proposed this concept in 2017, which refers to the use of artificial intelligence (AI) technologies such as machine learning and natural language processing (NLP) to automate the company’s raw data cleaning and analysis process. Turn data into actionable insights. Enhanced data analysis frees data scientists from tedious tasks such as cleaning data and manually analyzing problems, allowing enterprise users to obtain more detailed data insights in near real time.
Garnter predicts that enhanced data analysis will be fully popularized by 2022, and enhanced data management, which is closely related to enhanced data analysis, will also do a lot. Enhanced data management applies artificial intelligence and machine learning to all aspects of maintaining data quality, and establishes new methods to obtain value from existing data: for example, metadata that has been mainly used for auditing, research and reporting for a long time can be found in data Play a more active role in usage patterns.
Another representative technique is dialogue analysis. Gartner predicts that by 2021, this technology can increase the rate of employees (especially front-end employees) adopting data analysis and business intelligence from 35% to more than half. Conversation analysis combines natural language processing (NLP) and data analysis, allowing business users to ask questions naturally as if they were using search engines on the Internet. Then, the artificial intelligence algorithm analyzes the problem, runs the data analysis model, and feeds back the results again in a way that business users can understand.
The introduction of a large number of artificial intelligence technologies such as machine learning and natural language, as well as the increasing demand for real-time data insights in enterprise business scenarios, has created a demand for faster processing of larger data sets, and directly drives more and more companies to start turning Memory analysis platform to handle complex data analysis workloads.
A global survey by Forrester corroborates this megatrend.
2
Memory analysis platform accelerates the creation of intelligent enterprises
Nowadays, whether it is digital transformation or intelligent upgrade, the essence of enterprises is to obtain value from the massive and diverse data centers they have mastered, and truly realize "data-driven business" in the digital age. In this context, Forrester conducted a customized survey of 353 data management strategy decision makers in the United States, China, the United Kingdom, Germany and other countries in 2019.
The survey found that data has become a core factor affecting business success, and companies are accelerating their focus on real-time business intelligence; however, they are still facing a complex tool ecosystem that is not humane enough, making it extremely difficult to obtain value from data. However, through the use of persistent memory technology In-memory database solutions to build an in-memory analysis platform can effectively reduce complexity, eliminate data islands, simplify integration, and solve faults, allowing enterprises to accelerate their transition to "data-driven" intelligent enterprises.
For example, 61% of the interviewed companies encounter system limitations when collecting and analyzing big data. About 55% of the interviewed companies are unable to access data through self-service because the data access tools are not humane enough for the business. Due to the complexity of data and applications,% of the interviewed companies have difficulty integrating different data islands.
Enterprises and organizations now need to face a large number of multi-source, heterogeneous, and massive data environments. They will use a large number of data storage, access and integration solutions. In addition, data lifecycle management tools and processes are complex, and the complex tool ecosystem allows enterprises to Data management and analysis have become extremely complicated. In this survey, in view of the above-mentioned complexity, about 50% of the respondents said that it brought great difficulties and challenges to data management and analysis.
At the same time, the survey also shows that about 70% of companies have begun to implement in-memory databases to alleviate some of the above problems, and another 28% are planning to implement or are interested in the technology. Approximately half of the interviewees believe that memory capabilities can help artificial intelligence and machine learning applications, stream analysis and event processing, and gain real-time business insights from transaction systems. Real-time insights and intelligent business applications based on artificial intelligence and machine learning are one of the most important data analysis capabilities of enterprises and institutions.
As enterprises evolve into intelligent enterprises, they deploy more and more intelligent business applications, and the requirements for data processing and large-scale storage and retrieval of data continue to increase. The combination of SAP HANA memory database + Intel Optane persistent memory is considered to be Enterprises build an excellent combination of memory analysis platforms. This combination not only effectively improves data integrity and consistency, but also realizes real-time data conversion and calculation, thereby improving process efficiency and development speed.
In this excellent combination, Intel Optane persistent memory, as an important technological innovation, provides the best guarantee for "real-time analysis" in terms of performance, capacity, and reliability, and is also favored by more and more users .
3
Completely release the potential of the memory analysis platform
Why is "real-time analysis" important?
If you look at the current data growth situation, you can understand. By the end of 2020, there will be 5.8 billion Internet of Things terminal devices in use, and global mobile users have exceeded 8 billion. By the end of this year, the total global mobile data traffic will reach 38EB per month, and it is expected to increase to 160EB per month by 2025. . These trends actually indicate that an era of massive and rich data is accelerating. Obtaining insights in such a large amount of data in real time is extremely valuable to the business.
On the other hand, only by realizing the dynamic digital transformation of the core of the business can companies effectively deal with the impact of such real-time data. For this reason, it is necessary to instantly transform the huge information flow influx from various parts of the business process (suppliers, factory floors, distributors, and customers) into feasible data analysis, thereby improving competitiveness and efficiency.
For example, in Double Eleven, this year is the "Double Eleven" platform promotion + Internet celebrity marketing + live broadcast and other sales models. Although the "hot models" sell well, they may also encounter production capacity. Can’t keep up with the situation. It is also a small home appliance brand. Companies with data analysis capabilities often comprehensively analyze the capacity of each warehouse, product restrictions, upstream production capacity, transportation restrictions and other factors, and conduct comprehensive data analysis based on dynamic demand, so as to provide replenishment recommendations, realize allocation, Efficient in different scenarios such as distribution and distribution.
For many enterprises and institutions, the many in-memory database functions of SAP HANA platform provide core software functions required for real-time running of high-speed transaction processing and advanced analysis. It is one of the best choices for in-memory analysis platforms, but it is limited by the performance of existing hardware , So that its potential has not been fully realized.
The emergence of Intel Optane has brought a huge guarantee for releasing the potential of SAP HANA memory analysis. For example, HPE Superdome Flex set a world record of overall performance and 16 processor performance with 41.6 billion initial records in the SAP HANA standard application benchmark version of the SAP Business Warehouse version. The Superdome Flex used in the test was equipped with 16 second-generation Intel ® Xeon® Platinum 8280L processor, 12TB DRAM and 12TB Intel persistent memory. Compared with the DRAM-only solution, Intel Optane persistent memory can bring higher capacity and reduce the cost of SAP HANA platform infrastructure by 39%.
In fact, SAP HANA+Intel Optane is a combination favored by more and more users. For example, Evonik, a specialty chemical manufacturer, needs to cooperate with a large and complex supply chain to meet the needs of global customers. In order to track all logistics dynamics and deal with various challenges with ease, Evonik uses SAP HANA to perform real-time data analysis and generate reports. When using Intel Optane persistent memory for verification, the speed of reloading the data table after the server restarts increased by 17 times. Significantly shorten the SAP upgrade and maintenance cycle, and reduce the total cost of ownership.
It is often said that the world of martial arts is fast and unbreakable, and the value of "real-time" has been fully demonstrated in business scenarios in the digital age. As the pace of industrial digitization accelerates, digitization will cover all aspects of enterprises, and the enterprise that builds its memory analysis platform is like having a martial arts secret in the digital age. The faster this martial arts, the more differentiation in the market. In order to remain invincible.