Inspur software definition storage development concept is multi-application, promote technology development

As one of the development directions of future storage, software-defined storage has obvious advantages in capacity, performance and scalability. Performance and capacity are traditional enterprise users, but the combination of new technologies and business scenarios that customers pay more attention to. . In addition, with the continuous input of artificial intelligence technology, it also needs to be deeply integrated with different industry application scenarios, so that "smart" can reach the needs of users.

Software-defined storage for multi-services

Nowadays, enterprise data centers are transforming into clouds. How do data cloud infrastructures on public and private clouds be built? How to ensure the smooth operation of the disaster recovery backup of the three centers in the two places? How to achieve uniform storage of various storage types such as SAN storage, object storage, and NAS storage? How do the underlying architecture and services adapt to the rapid changes in the scene? These all require software-defined storage to innovate in changing market demands.

Data is not treated “equivalently” to users. Some of the hot data in the storage system will slowly cool down and eventually be deleted or migrated to low-cost, high-capacity devices. All data is unrealistic in high-end storage. Users need a platform to allow flexible data flow. An interface can manage heat, cold and temperature data in a unified manner, flexibly dispatch, and master the life cycle of data. This is also intelligent storage. The essence.

[Inspur Software Defines Stored Focused Scenes for Deep Optimization]

Today, Inspur Intelligent Storage integrates more than 1000 hardware and software sensors, 100+ core algorithms for continuous optimization, “intellisense, intelligent architecture, intelligent management” – emphasizing intelligent management of data lifecycle, multi-scene application awareness, capacity adjustment, Intelligent features such as performance acceleration, reliable hardening, heterogeneous integration, and data migration.

In 1 year, Inspur launched the industry's first "four-in-one" whole cabinet software definition storage AS13000-Rack to undertake video surveillance, radio and television media, artificial intelligence, cloud computing, big data, high performance, database seven scenes and hundreds A secondary application scenario, storage service + software + hardware three-tier architecture, software and hardware collaborative optimization, providing a total solution for SAN, NAS, object storage and big data application, centralized customization for each scenario and each application Processing, 100,000 IOPS, bandwidth up to 2.5GB / s, single-node performance leading.

Zhang Ke, senior product manager of Inspur Storage, pointed out that the application of artificial intelligence and the growth of massive data have placed high demands on the performance and scalability of traditional storage. Therefore, software-defined storage will be more in the future. Deploy in the data center.

Inspur software defines four optimizations for storage

Zhang Ke said that Inspur has completely optimized the software-defined storage under the scenario to help enterprises solve problems in practical applications.

First, performance acceleration based on file IO aggregation. Video surveillance involves a large number of high-performance small file processing, while small files have high requirements for metadata and data performance during the reading process. The wave is aggregated by small files in the software layer (or split large files). Aggregate 4K and 8K small video into 4M data packets, and aggregate the IO of the client application layer without affecting the underlying storage performance, greatly improving the IO of the service and reducing the performance bottleneck caused by the IO pair storage access. To meet the needs of large-scale storage applications, greatly reducing the high concurrency of large and small files, and improving the performance of client applications by 5 times. It is widely used in traffic jams, animation rendering and geological exploration.

Second, unify multiple storage protocols. Inspur supports NAS, SAN, Object and HDFS protocols and seamlessly connects to practical applications, reducing performance loss during data copying. Data is converted between the NAS and the object protocol at the software layer, and the data can be processed uniformly. This is mostly used in meteorological education, aerospace and other operations. Zero-copy of multi-source data eliminates the silo of storage information, and a set of storage can satisfy a variety of applications of customers, which not only saves hardware cost procurement costs, but also improves management and operation cost efficiency.

Third, WAN-based remote replication acceleration. Mainly used in the case of high latency or high packet loss rate, from the original single data source transmission to multiple virtual channels, complete hundreds, thousands of kilometers of data replication, bandwidth utilization increased by over 90%, equivalent to ten groups Data transmission at the same time, any group of transmission interruption or data loss, other channels are not affected, disaster tolerance capability is more perfect than RAID, and the security and efficiency of long-distance data transmission are high in satellite transmission and off-site data centers. This is especially true for scenarios.

[Inspur Software Definition Storage]

Fourth, the stream storage technology that enhances the performance of video surveillance storage. The traditional video surveillance solution is mixed with the requirements of the management software. The data is transmitted from the front-end camera to the management platform and then transmitted to the client. The AS13000 directly transfers the data transmitted by the camera front-end GPU to the storage, and the streaming storage and storage directly saves the hardware. The streaming media server and the management platform save the cost of the media service hardware facilities; on the other hand, the streaming media server has a single point of failure, and all the data is directly stored on the software-defined storage device through the camera, and is guaranteed by the erasure code redundancy mode. Data security; in the face of massive video data, the system bears the test of video capacity. Delay and bandwidth will seriously affect the work efficiency. The wave increases the link bandwidth capacity, improves network QoS, and responds to massive data storage.

Soft and hard collaboration, service multi-application

The development concept of Inspur software definition storage is multi-application and promote technology development. Around this core, the next step, for user QoS, Inspur will customize the capacity and performance according to different business scenarios of users; at the same time, ensure hardware stability, soft and hard coordination, and accelerate performance improvement; based on hardware reconfiguration, strengthen software layer Function, do a good job of docking cloud data; In addition, Inspur will do distributed full-production support on software-defined storage in the next year, and work with partners to integrate technology in a multi-application scenario. In the future, artificial intelligence will be more used by the wave for deep optimization of software, including the old equipment management in the background can also be easily solved through a set of software, the wave is working hard for this.

Can be expected,

2018 will be a year of software-defined storage scale applications!