Getting Ready for the Yottabyte Era: Decoupling Computing and Storage
The production of data in the digital world continues to accelerate. AI and edge computing in particular are contributing to exponential growth. In order to productively use the enormous amounts of data generated, they must be captured, stored, transported and analyzed. This in turn requires enormous amounts of computing power and storage. Separating computing and storage capacities is a solution approach to this challenge that leads to greater efficiency and cost savings.
The global data volume is unimaginably large[i]. According to a study by market researcher IDC, the global volume of data will be ten times greater in 2030 than it was in 2020[ii]. By then, it will amount to around 570 zettabytes. Other experts predict even more growth: According to a study by Huawei, one yottabyte of data will be generated worldwide every year by 2030, which would correspond to a 23-fold increase compared to 2020[iii]. In case this doesn't mean anything to you: 1000 gigabytes make a terabyte, 1000 terabytes make a petabyte; an exabyte consists of 1000 petabytes, a zettabyte consists of 1000 exabytes, and a yottabyte counts 1000 zettabytes[iv]. In the light of such figures, efficient data processing and data storage are all the more important, not least because the utilization of such data consumes huge amounts of energy.
The Yottabyte era begins
The storage industry is therefore facing major challenges because storage technologies have to keep pace with the ever faster flow of data. After all, this data stream is digital gold if it can be harnessed. However, not only the mining of relevant data, but also its identification and efficient use is not entirely straightforward. Since Big Data systems demand a lot of computing and storage resources, new capacities have to be added again and again as the data stream swells. This is where the cloud can play an important role - in connection with the separation of storage and computing capacities.
Telecommunications group and storage specialist Huawei is developing a new storage concept for such use cases. This involves “building a data-centric, trusted storage foundation for a wide variety of applications”, as outlined at the 2022 Innovative Data Infrastructure Forum. Therefore, the company will prioritize the development of decoupled storage-computing architectures and the acceleration of data applications.
The decoupling of storage and computer
To achieve the goal, progress is needed in three areas, notes Dr. Peter Zhou, head of Huawei's IT product line.
As enterprises continuously accelerate their digital transformation, new data applications are increasingly becoming production applications that require higher data reliability. In addition, the difference in the lifecycle length of data and computers is widening, requiring flexible and independent management and maintenance of computing and storage resources. This calls for a decoupled storage-computing architecture to better leverage elastic and low-cost storage and enable faster and more reliable services at a lower cost.
The acceleration of data applications
According to Huawei, storage systems need more components to accelerate data applications. There are more and more business applications generating massive amounts of data, for example, decentralized databases, Big Data, and AI applications. Future storage systems should therefore be equipped for metadata management and intensive data processing to form various data acceleration modules that could increase processing efficiency by up to ten times and significantly improve customer experience and operational efficiency.
The greening of design
Finally, according to Huawei's head of products, a green design system needs to be created to fundamentally improve the product development process. Various aspects need to be considered, for example, renewable materials, low-carbon components, such as NAND flash, and high-density hardware. Supporting business services that optimize production operations to achieve greater energy efficiency should also be included.
The development of new computing and storage systems for the massive data flood of the future is therefore underway. In Switzerland, UMB is a Huawei Gold Partner, has broad Huawei know-how and has already carried out numerous successful customer projects. As a Gold Partner, UMB is also a Certified Service Partner and can provide direct service locally for Huawei customers.
As for Huawei storage systems, UMB is currently focusing on OceanStor Dorado V6 NVMe all-flash storage systems, which are tailored to a wide range of enterprise sizes and provide superior storage performance at a lower total cost of ownership. Please contact us for additional information.