The D-Fusion SOC series is a large-scale distributed cluster storage tailored for high-end users. It is based on a scalable cluster architecture to meet the expansion needs of enterprise storage capacity and storage performance. Expanded industry-leading, enabling the expansion of more than 256 storage node groups; as storage node groups increase, performance grows linearly, storage capacity can grow to 160PB, providing more than 1024Gb/s bandwidth and 25 million OPS。
        The D-Fusion SOC series consists of multiple storage node groups. Each storage node group is redundantly designed and consists of two active storage controllers (Active-Active). Nodes connect and exchange information through IB. During the operation of the system, any storage node group failure will not affect the normal operation of other storage node groups, and there is no single point of failure, which improves the service level of the system. The modular design of the system provides us with usability and flexibility in addition to superior performance and scalability. At the same time, the D-Fusion SOC series reduces management and staff costs, reduces operating expenses, and provides high-quality integrated services by integrating hardware reduction and simplified management.。
        The D-Fusion SOC series uses NAS to support storage for front-end systems. Through the NAS access protocol, CIFS (SMB), NFS can provide file system access: NAS clients can use a single NFS mount point or CIFS share to access huge data containers. Each client only needs to install an NFS file system mount point or access a CIFS share. Within the Prodigy OS provided by the D-Fusion SOC series, the file system is managed by the administrator. For NAS clients, each file system appears as a folder or subdirectory nested under the NFS file system mount point or the root of the CIFS share. The file system can be added at any time, and the client can use the added file system immediately without having to see the new storage and reinstall it. In addition, because the total file system capacity is not fixed, multiple users can share a storage pool with a unified capacity.。

The D-Fusion SOC series includes the following products:
        1) D-Fusion 8000 SOCseries：Ultra-high-end performance unified horizontal distributed cluster storage system, supporting 256 storage node group clusters, dedicated to high transaction type and large throughput-intensive applications. The D-Fusion 8000 SOC Series delivers high-performance bandwidth and OPS in existing scale-out clustering solutions. The unique hash algorithm enables load balancing on the storage side, leveraging storage resources and improving storage node group utilization. The D-Fusion 8000 SOC features a high-performance Infiniband host port with an enterprise-class 15k rpm disk and high-end global cache protection. With the support of the PRODIGY operating system, the performance of the entire storage system is fully realized.。
        2) D-Fusion 5000 SOCseries：A horizontally distributed cluster storage system with a high comprehensive index for enterprise-critical business supports 128 storage node group clusters to achieve a balance between high capacity, high performance, and low cost. The D-Fusion 5000 SOC series can use PCIe technology to play the second-level cache, which greatly speeds up the reading and writing rate of the entire storage system, and the data processing capability is rapidly improved. It is a horizontal distributed cluster with high throughput and high concurrent applications. Storage System。
        3) D-Fusion 3000 SOCseries：Cost-effective, highly scalable mid-range horizontal distributed cluster storage system supporting eight storage node group clusters. Whether used as primary storage or secondary storage, the D-Fusion 3000 SOC Series delivers exceptional versatility and connectivity to meet your data protection and data management needs at an affordable cost, increasing storage efficiency。

System advantage
Scale out
       Scale-out is a way to address the growing demand in the storage environment. The scalability of all storage controllers is limited by physical conditions: the number of CPUs and memory slots, and the space on the disk shelves. These conditions determine the large capacity and performance that the controller can provide. If you need more storage capacity or performance capacity, the general approach is to add CPU and memory, or install more disk shelves, but eventually the controller will be completely filled, and you cant add hardware. At this point, the first option that comes to mind is to add another controller. One way to do this is“Vertical expansion”：That is, each additional controller is a completely separate management entity and does not provide any shared storage resources. If you want to completely replace the original controller with a newer controller, you need data migration to migrate data from the old controller to the new controller. This work is time consuming and potentially damaging, and it is possible to make changes to the configuration of all host systems to be connected.。
        If the new controller is to coexist with the original controller, there will now be two separately managed storage controllers and there is no built-in tool to balance or redistribute the workload between the two. This situation gets worse as the number of controllers increases. By using upscaling, the operational burden increases as the environment grows, and eventually a very unbalanced and unmanageable environment emerges. The technology refresh cycle requires careful planning ahead, and long-term outages and configuration changes can pose risks to the system.。
        In contrast, using scale-out means that as the storage environment grows, other controllers are seamlessly added to the resource pool on the shared storage infrastructure. Host and client connections and data repositories move seamlessly and non-disruptively across resource pools, making it easy to balance the existing workload of available resources while also easily deploying new workloads. Complete technical updates (replace disk shelves, add or completely replace storage controllers) while keeping the environment online and providing data。

 high performance
        The D-Fusion SOC family is the first to achieve scalability of up to 256 storage node groups, consolidating more file servers and application storage on a single unified storage platform, and its performance scales nearly linearly with the expansion of storage node groups, storage Capacity up to 160PB。
       The D-Fusion SOC series meets the extremely high performance requirements of high-performance computing clusters. With the increase of storage node groups, the bandwidth can be as high as 1024Gb/s, and the OPS can be as high as 25 million. It is suitable for thousands of nodes and extremely high performance requirements. Computational cluster. At the same time, the D-Fusion SOC series supports intelligent load balancing and dynamic thin provisioning, enabling enterprises to flexibly expand storage as business needs change.。

High scalability
        The D-Fusion SOC series inherits the vertical expansion technology of D-Fusion Standard Edition products, realizes the integration of Scale-up and Scale out technology, and breaks through vertical expansion cost-effectively to achieve all-round expansion. When the D-Fusion SOC series flexibly expands, it can not interfere with business processes and be transparent to the client, thus ensuring the sustainable development of the business side.。

High storage efficiency
        The D-Fusion SOC Series delivers industry-leading advanced storage efficiency, data management and data protection。
        Intelligent distributed storage cluster management, D-Fusion SOC series realizes efficient horizontal expansion with the power of PRODIGY operating system。
        Efficient IB switching technology, D-Fusion SOC series products exchange data through IB internally, realize fast data transmission between storage node groups, and combine the unique hash algorithm to return data to the client, thus making data access fast and transparent。
        Manage, manage smart, simple and convenient with Prodigy Viewer management tools。
        Security rights management, virus scanning, static data encryption and other functions integrate data security into the D-Fusion SOC series, realizing unified and streamlined data management and protection.。
        The D-Fusion SOC series can adopt deduplication technology to achieve high data compression rate, thereby improving storage system resource utilization and reducing costs.。

High reliability
        The reliability of the D-Fusion SOC storage system exceeds 99.999%, which means that the annual downtime (including planned downtime) is less than 3 minutes. It provides high reliability assurance measures from hardware, system and architecture.。
Hardware layer
        All hardware components of the D-Fusion SOC storage system, including controllers, power supplies, fans, and network connections, are redundantly designed。
       The D-Fusion SOC storage system uses real-time data checksum to automatically check and repair potentially corrupted sectors of the disk to ensure data integrity.。
System layer
        The D-Fusion SOC storage system provides multiple redundancy modes such as RAID 0, 1, 5, 6, 10, 50, 60, and triple parity. It supports hot spare disks and automatic RAID reconstruction. D-Fusion SOC storage systems are created faster, usually no more than 10 seconds. RAID 6 prevents data loss or corruption in the event of simultaneous damage to both disks. D-Fusion SOC storage systems unique dynamic striping RAID technology can also effectively prevent RAID“Write vulnerability”Phenomenon occurs;
        The PRODIGY-EP file system of the D-Fusion SOC storage system is very robust, maintains consistency in extreme cases, and automatically recovers quickly when the system is rebooted. The D-Fusion SOC storage system does not have a separate metadata server. All metadata is managed by the controller and distributed on disk to obtain the same reliability protection as the data. This approach is far more reliable than a separate metadata server, because the latter interrupts the business when the metadata server crashes, and it takes a long time to complete a file system consistency check task like fsck, and more serious because of metadata. Get full recovery and cause some or all of the data to be corrupted or lost。
Architecture layer
        Multiple D-Fusion SOC storage controllers are connected to each other to form a high-availability cluster. The cluster working mode is Active-Active, that is, all controllers can provide data access at the same time, and the traditional active/standby working mode (Active- Compared to Standby), the performance of data storage is significantly improved. At the same time, the two or more controllers are hot standby, and any one of them fails, and the other can automatically and continuously take over the task of the fault controller and the data managed, and this is completely taken over. The process is completely transparent to the client. This guarantees 7x24 business continuity. Other manufacturers“NAS Gateway + SAN Array”Most of the structures can only be backed up by SAN arrays, and cannot be backed up by NAS gateways.。