AT GITEX this year, tech behemoth Huawei launched DataLake which follows the ‘One Cloud, One Lake, One Platform’ architecture in an effort to help customers innovate their services and increase the value of their businesses based on application scenarios of customers from sectors such as public utilities, finance, transport and manufacturing.
Smart Cities World Forums sat down with An Jian, Huawei Middle East’s President of Carrier Networks, to get more insight into the DataLake launch and to discuss the importance of data in today’s world.
Huawei's storage product line has been renamed as the intelligent data and storage product line. What is the background of this decision?
In this era of big data, Schonberg wrote, "Data can quantify everything. Words become data, orientation becomes data, communication becomes data, and everything is data." From this perspective, data seems to be the main theme of the modern world.
Data is an intangible asset for companies worldwide. Carriers are the leaders of digital transformation, and the advent of 5G will make carriers the drivers of digitalization for all of society. Converting intangible data into tangible productivity is what Huawei, together with carriers, strives for. To achieve this, Huawei integrates data-related products and solutions based on the storage product line to meet the needs of the data life cycle, simplifying data storage and maximizing data value.
You just mentioned that Huawei will pay more attention to data storage and data value in the 5G era. What challenges do you think exist in relation to this?
According to the Huawei Global Industry Vision (GIV) report, the global volume of data will see an enormous increase from 32.5ZB in 2018 to 180ZB in 2025.
In addition, data will become a critical production factor in the 5G era. Data production and mobility are accelerated by 5G, 4K/8K videos, IoT, V2X, and other new apps. However, less than 2% of this is stored, and less than 10% was analyzed and utilized. Today, when we mention data, it is clear that how to store and analyze data correctly are enormous challenges.
The first challenge is that carrier data has been stored in the silos of the production systems, resource utilization rate is less than 30%, power consumption and space costs have reached millions of dollars, and the TCO cannot be effectively reduced. In addition, data storage requirements cannot be satisfied, capacity expansion takes a senior administrator more than 6 hours, and the silo architecture cannot support customers’ cloud transformation or future business development.
The second one is being able to efficiently extract useful information from massive data and explore data value is another huge challenge facing carriers. During this process, data must go through the three stages of access, analysis, and consumption. In the data access phase, customers are usually met with the challenge of how to connect tens of thousands of access points and convert different protocols.
In the data analysis phase, customers have to solve the issue of how to integrate data from different domains for fusion analysis, and how to locate target data from among millions of tables. In the data consumption phase, customers are met with the problem of how to shorten the business development process to less than one day. These are the challenges facing the analytics domain.
How can Huawei help carriers address these challenges?
After studying numerous carriers, we proposed a 5G data network solution. This solution includes two parts: an intelligent storage management platform for the production domain, and an intelligent data platform for the analysis domain.
In the production domain, Huawei's intelligent storage management platform OceanStor DJ integrates multiple types of storage systems, such as Huawei's all-flash storage, distributed storage, and storage from third-party vendors.
It then consolidates the different resources and different types of storage into a unified resource pool, automatically distributes them according to different business SLAs, and implements on-demand, flexible scheduling of storage resources. Based on this, it develops the ability to undertake automated planning, allocation, O&M and optimization of the full life cycle. This helps customers significantly reduce storage TCO, speed up TTM, greatly improve the ease of using storage, and achieve evolution to future storage architecture.
In the analysis domain, we must overcome the challenges of difficult data access, difficult analysis, and difficult consumption, open up global data connections, build a unified data platform, and improve real-time data service capabilities.
Huawei's intelligent data solution includes three components: data access, data processing, and data enablement. Data access solves the issues with diverse data access and connects all carrier applications and data. Data processing overcomes the problems related to data fusion analysis by realizing efficient data analysis without the need for data relocation.
Data enablement solves data consumption problems, enables users to enjoy the data self-acquisition-and-analysis experience, and improves data usage efficiency. In addition, it makes business more agile.
What kind of value can Huawei's 5G data network bring to carriers?
Based on actual usage, Huawei's 5G data network solution can increase storage space utilization in the production domain from its current level at less than 30% up to 70% and reduce TCO by more than 30%. The solution can also enable storage resources to be applied by the service demander, and then allocated automatically.
This reduces the average resource allocation time from 6 hours to 5 minutes, a decrease of more than 90%. In this way, the difficulties with storage system management are greatly reduced, and the management cost can be reduced by more than 50%, accordingly. In the analysis domain, Huawei's 5G target network solution has been used on a large scale in Huawei internal IT systems. This solution helped Huawei reduce each ETL task, including development and implementation, from two Human to one Human days, enabling data to be quickly shared among multiple analysis systems, improving cross-regional data access performance by multiple times, and reducing the time of developing data consumption applications from three days to one day.