Five current trends in data storage

Billions of IoT devices are generating ever greater volumes of data. This data holds enormous potential; among other things, it can be used to develop new business models or make manufacturing processes more efficient. But the new appetite for information also brings challenges: How can the floods of data be stored efficiently? What are the future trends in storage? The industry is constantly trying to answer these and other questions. Here are the five trends that are currently important in the field of data storage:

Shift from cloud to edge computing

In the early stages of IoT, applications were primarily supported by clocud platforms. However, computing power and intelligence are increasingly shifting from the cloud to the network edge. The interplay of 5G, AI and IoT makes it possible to process significantly more data directly there. For many IoT applications, it is now essential that data is verified, analyzed, and transformed at the point of collection, for example in vehicle sensors, robots, or drones. This creates more demand for storage in edge devices, which in turn creates new challenges. For example, the highest possible, but above all scalable, storage capacity is needed at the network edge. Especially when a large number of endpoints are connected to each other. How the storage solution has to be designed also depends on the environment, for example. The more exposed or even inhospitable the location where the data collection and processing takes place, the more special the requirements for durability and reliability of the storage.

Network latency plays a key role in IoT applications. If processing is too slow, the captured data can be corrupted or even lost. Edge storage is capable of managing data collection and provides sufficient compute resources to aggregate and analyze this data in real time - for example, through the use of AI or machine learning (ML). This creates immediately usable insights at the device level.

New preferred memory type for IoT applications

Reliable and fast memory solutions are critical for optimal edge device performance. NOR and NAND flash memory are typically used in IoT applications. NOR is stable as well as reliable, and features include long lifetime, low power consumption, and execute-in-place (XIP) capability for faster boot-up. NAND, on the other hand, is faster in every other respect, takes up less space, is less expensive and achieves higher density.

Traditionally, NOR was used as code memory with high reliability, while NAND was used for larger capacity and higher performance. In recent years, this has changed. NAND is also becoming more popular for code storage. This paradigm shift will continue to grow as IoT applications become more sophisticated and modern edge devices can store code. NAND-based storage solutions such as eMMC (Embedded Multimedia Card), UFS (Universal Flash Storage) or SSD (Solid State Drive) are coming to the fore in almost all sectors - from the manufacturing industry to medical technology.

Cold Storage Increases in Importance

In a changing - increasingly hybrid - working world, the requirements for storage solutions are also changing. When colleagues collaborate from multiple locations, critical data can no longer be stored in just one central location. To still be able to access and protect the data, companies need to use local, external and cloud backups as well as archiving solutions.

However, multiple backups at different locations significantly increase the amount of storage in data centers. It is therefore important to examine how data can be archived most efficiently. Long-term archives, also known as cold storage, offer a suitable solution. Data that does not need to be accessed immediately is stored cost-effectively and archived until it is needed. The increasing importance of cold storage is also a response to cyber attacks such as ransomware attacks. These are much less dangerous if external backup copies exist.

Most cold storage archives are housed either on tapes or on hard disk drives (HDDs). Tape storage is less expensive than HDDs, but also has higher latency when accessing data. HDDs and next-generation platforms have a critical role to play in further increasing the efficiency of archive solutions.

(Energy) efficient data storage is essential

In order to cope with the ever-increasing flood of data, many companies use data centers. However, these consume huge amounts of electricity and thus account for a not inconsiderable share of global CO2 emissions. Relying on energy-saving data centers not only pays off ecologically, but also financially and lowers the total cost of ownership. There are various approaches to increasing energy efficiency in data storage.

Thus, it makes sense to switch to high-capacity drives. A few high-capacity drives are more energy efficient than multiple low-capacity drives. At the same time, this increases the storage density per rack, which in turn reduces their required number. JBoD (Just a Bunch of Disks) solutions can also reduce power consumption. Here, the drives are protected from vibrations using modern technologies, such as those we offer exclusively, and the airflow is channeled in such a way that optimal operating temperatures can be maintained. Additional expensive cooling is no longer required. In this context, it is also worthwhile to use helium-filled hard disks instead of air-filled ones. Since helium has only one seventh the density of air, there is less turbulence inside the drive.

Ultimately, the recipe for success is to modernize the existing infrastructure and make it more efficient. There are several new approaches that help with this. For example, zoned storage is a good tool for storing data intelligently and thus increasing performance and storage density, reducing latency and lowering operating costs.

Success with an intelligent storage strategy

The examples make it clear that storage solutions are a critical factor in the use of IoT on many levels, but also for the fundamental performance of companies. An intelligent storage strategy is therefore indispensable and should be considered early on in all investments.

The ever-increasing volume of data streams is creating new challenges in storage. However, with the right solutions, these can be mastered.

 

NOR vs. NAND Flash

NOR flash is usually used for the program memory of microcontrollers and is also used, for example, as non-volatile memory for the BIOS or UEFI of a PC. Due to the internal serial circuitry in NAND flash, reading and writing is only possible in blocks.