Challenge 1: Storage Technologies That Deliver Ultimate Per-Bit Cost Efficiency

With the emergence and development of new non-volatile media and high-speed network protocols, further research needs to be done on storage architecture, media application, and storage media to build a storage system that delivers ultimate per-bit cost efficiency.

Direction 1: Storage Architecture Innovation

1. New storage-compute disaggregation technology: Research high-performance, high-reliability pooled sharing and elastic scaling of memory and storage resources based on the new high-speed network bus (compute express link/unified bus); and build tiered memory storage systems to implement diskless and transparent resource pooling on the host side and to provide memory-level data processing efficiency, local-site like performance, and a superior user experience.

2. New storage system architecture: Explore data-control separated, full-switching storage architecture, as well as controller-media separated/composable storage architecture. These architectures allow the combination of computing resources and computational storage capabilities (to implement near-data processing) which reduces unnecessary data migration.

3. Network-storage-computing convergence technology: Integrate network elements such as DPUs and switches with storage semantics, and develop the inline acceleration and computing capabilities of file systems, blocks, objects, and vectors; and build efficient data mobility, cache, and consistency protocols for network layers both within and between data centers to provide smooth cross-site access for users.

Direction 2: Media Application Innovation and Storage Media Breakthrough

1. New media that support mass production: Develop new storage media with low costs and fast access to break through the constraints of existing media technologies. The new media need to be able to be mass-produced.

2. Large-capacity disk technology: Reduce the power consumption of larger-capacity disks using advanced encapsulation and power consumption control technologies to deliver better single-disk capacity and performance simultaneously; and use technologies such as memory tiering and new indexing to reduce the memory consumption of large-disks scenario and the system without affecting system performance.

3. New disk form: Develop more resilient and reliable disk forms and disk interfaces that are suitable for the emerging scenarios of vehicle-mounting and edge; promote the separate, on-demand deployment of disk controllers and media to allow users to customize the combination of disk capacity, performance, and reliability so that they minimize the impact of media faults; and develop computational disks with scenario-based acceleration capabilities that are suitable for AI and big data in order to significantly improve data throughput and energy efficiency.

4. Disk/Media reliability technology: Improve the reliability of disks that have increased capacities to reduce the disk/partial failure rate, optimize data recovery speed, and reduce the capacity loss caused by media faults; and enhance the reliability of multi-media hybrid storage systems, such as magnetoelectric storage systems, to reduce the failure rate and fault impact duration even though the increased disk capacity/type may be pushing them up.