1         End-to-end 10Gbit/s architecture, providing world-class storage access

At the front-end, the Neocean IX3620 provides four 10Gbit/s host interfaces and at the back-end up to 72Gbit/s disk access bandwidth. Based on a number of 40Gbit/s PCI-E buses, the system provides an end-to-end non-blocking 10Gbit/s architecture.

l           10Gbit/s hierarchical front-end storage access network

3400MB/s transmission bandwidth: The IX3620 controller comes with a purpose-built 10Gbit/s TCP/IP Offload Engine (TOE) processing module to deliver high-bandwidth and low latency data access. The IX3620 provides four 10Gbit/s host interfaces capable of data transmission at 3400 Mbit/s. The controller provides industry-leading I/O capabilities.

10Gbit/s performance: The IX3620 host access network is built on 10Gbit/s Ethernet. With a 10Gbit/s TOE card, the front-end host can provide storage connectivity to the IX3620 for superior access performance in support of 10Gbit/s internetworking and suit bandwidth-hungry storage applications.

Hierarchical and multi-layer storage network: The front-end host can also come with a 1G network adapter or iSCSI host-bus adapters (HBA) card to provide 1Gbit/s centralized access capabilities at the host and 10Gbit/s port convergence capabilities at IX3620 storage side. This storage network revolutionizes the single-rate legacy SAN networking architecture. As well as significantly enhancing the storage access to handle server access on a larger scale, the networking architecture helps streamline the storage network structure, and when combined with QoS and security technologies, greatly enhances the manageability and scalability of the storage network.

l           Full switched and high bandwidth back-end disk access channel

The limited bandwidth for back-end disk access usually bottlenecks the entire storage system. Built on the cutting-edge serial-attached SCSI (SAS) bus technology, each SAS bus can provide up to 12Gbit/s access bandwidth, significantly enhancing the performance of the back-end disk access channel. An advanced bus technology based on the serial communication protocol infrastructure, SAS incorporates the advantages of existing parallel SCSI and serial connection capability. By utilizing the SCSI-3 extended instruction set and providing compatibility with SATA devices, the technology is a multi-layer connection protocol for storage devices.

72Gbit/s disk swapping bandwidth: At the back-end, the IX3620 uses the SAS channel as the data exchange link between the controller and disks. Every SAS channel between the controller and disk cabinet is designed to have wide ports. Each individual SAS wide port provides a data switching bandwidth up to 12Gbit/s. The system utilizes up to six SAS wide port disk channels with data switching bandwidth of up to 72Gbit/s.

Full switched intelligent disk cabinet: Built on SAS switching technology, the IX3620 disk cabinet provides an internal switching throughput of up to 144 Gbit/s. Every disk has separate data access links independent of other disks. This helps guard against performance deterioration as a result of shared bandwidth based on a ring architecture.

l           Leading-edge disk technology

The IX3620 supports a mixture of SAS disks and SATA disks. Based on dual-port and full duplex serial transmission, SAS disks have enhanced data transmission capability while providing high reliability and manageability usually built into legacy SCSI disks. The SAS disk used in the IX3620 system operates at 15,000 revolutions per minute (r.p.m.), with a Mean Time Between Failures (MTBF) of up to 1.4 million hours. The disk is suited to access-intensive, high-performance applications, especially those that require supercharged random access capability. The SATA disk used in the IX3620 system also provides an MTBF of up to 1.2 million hours, significantly reducing the investment and total cost of ownership (TCO) while delivering high reliability for users.

l           High-performance multi-core storage controller

The IX3620 storage controller integrates in its design multi-core processor, PCI-E bus, FBD memory and other world-leading data processing technologies. It ensures the high I/O performance and service expansion capability of the controller.

Multi-core processor: The IX3620 adopts the Intel high-performance multi-core processor. Based on the 65nm process, the processor adopts the “dual independent bus” and supports the new generation processor virtualization technology (VT). The processing performance is three times that of the single-core processor of the same frequency. The IX3620 integrates four dual-core processors, and provides faster, more reliable storage capability, stronger virtualization capability and faster IO processing capability.

PCI-E high-speed bus: To manage the internal interconnection of the controller, the IX3620 adopts the PCI-E bus interface standard. PCI-E utilizes industry-leading point-to-point serial transmission and each bus provides 40Gbit/s access bandwidth. Each set of the equipment has its own dedicated connection, without the need of requesting bandwidth from the entire bus – significantly speeding data transmission rates. Internally, the storage controller adopts multiple high-speed PCI-E buses, providing a 10G data switching bond between the multi-core processor, 10G host channel, and SAS disk channel. This approach ensures end-to-end, non-blocking 10G data transmission capabilities.

In addition, the IX3620 storage controller adopts the new generation serial transmission FBD memory. It supports the multi-channel simultaneous read/write of the memory, and provides memory access speed faster than DDR as well as better expansion capability. The device integrates an independent hardware RAID processing module. While providing efficient disk access, it greatly reduces the computing load of the processor.

2          All-dimensional high availability design to ensure data reliability

The Neocean IX3620 adopts multiple available technologies to provide complete system fault protection, ensuring maximum availability and security of users’ data.

l           Multi-path redundancy and controller load balancing

The IX3620 storage controller integrates multi-path redundancy and intelligent load balancing functions. It provides end-to-end service sharing and redundant protection to entities from application servers to storage arrays. When the host access link, disk access channel, or storage controller is faulty, the system can avoid single-point fault, ensuring continuity of storage service. More importantly, regardless of path switching or controller switching, the entire process does not need participation of the application host. As a result, the impact on applications is reduced maximally.

l           Write cache mirroring and data protection

The IX3620 adopts write cache mirroring technology to write the same data into the caches of two controllers at the same time. This approach avoids cache data loss due to the failure of a single controller. In addition, the system is configured with the backup battery protection for the write cache. It can provide up to 100 hours power-off data protection.

l           Redundant power supply protection

The main components of the IX3620 have a redundant power supply design, providing higher reliability. When one of the power inputs is faulty, the other power input can still support normal operation of the system. The storage controller at the core position specially adopts the 2+2 redundant power supply protection. In addition, the power supply supports hot swapping and online replacement. It also provides protection to ensure equipment security in cases of output overcurrent, output short circuit, and output overvoltage.

3         Diverse software functions and a large variety of solutions

The IX3620 provides diverse data management functions, for example, fully automatic continuous data snapshot, network duplication, snapshot agent, and DiskSafe. It implements data protection, from online to near-line and local to remote. It offers multi-layered, cross-regional storage solutions to the users.

Fully automatic continuous snapshot: The IX3620 supports the TimeMark fully automatic continuous snapshot function. It can create automatically increment-based, multi-time point marks (up to 255) according to polices. The time point mark policies can be created on a need basis. For example, the creation can be done at 12 o’clock in the evening, in every other hour, or in every other ten minutes. When data damage, virus damage caused by software program, accidental deletion and data loss, or error caused by human misoperation, the “rollback” operation can be performed to quickly restore data through proper time point mark. The function is particularly applicable to the continuous data protection of key services.

TimeView (provided together with TimeMark): TimeView can capture data volume at any time point state and provide it to the host as an independent data volume for data test, data analysis and data backup. TimeView is an important tool for searching for the correct time point data when the production volume is faulty. Each time point marked by TimeMark can be extracted as data volume for analysis.

Data snapshot replication (provided together with TimeMark): The IX3620 series also includes a data snapshot replication (data cloning or volume replication) function. Snapshot replication refers to the physical “cloning” of the source data volume into a completely identical data volume. The cloned volume and source data volume are two fully independent, physical volumes. Any damage of the original data volume will not affect the cloned volume.

Network replication: The policy-based replication function of the IX3620 can implement non-local backup of user data. In the case of an accidental disaster, fast recovery of the data can be implemented, ensuring the continuity of user service.

The internal system of the IX3620 will perform virtual division of the specified volume intelligently into data blocks, minimally 512 bytes. When the application keeps modifying the data, the IX3620 will intelligently identify the changed data blocks. It will replicate only the changed data blocks instead of the entire file, table or directory. This approach enables efficient implementation of the replication within the shortest time and least bandwidth.

The remote replication function of the IX3620 supports three policy replication modes based on time interval, data increment and preset time point. Flexible setting can be done according to the application requirements. The replication technology of the IX3620 can provide encryption and compression options. It can perform special processing of service data, ensuring data security and maximized utility of bandwidth.

In addition, the IX3620 includes an auto-sensing replication function. It can replicate in real time each IO written in the main resource to the copy resource. As a result, the amount of data loss in the case of disaster is minimized.

Host data consistency (snapshot agent): When the host application is operating normally, part of the data has not been written to the back-end disk. Instead, the data stays in the memory of the host. If snapshot operation is only performed for the data in the disk, inconsistency between snapshot data and expected data occurs. The IX3620 also provides a snapshot agent function to users. After a user sets the TimeMark and replication policies, the IX3620 will automatically interact with the host snapshot agent when the IX3620 executes the operations according to instructions. It will then automatically inform the snapshot agent to copy the data that temporarily stays on the memory of the application to the disk. This measure ensures the strict data consistency and completeness of the snapshot or replicated data.

The snapshot agent provides support to various platforms, including applications such as SQL Server, Oracle, Sybase, DB2, Informix, Exchange and Domino and operating systems including Windows, Linux, IBM AIX, HP-UX and Solaris - as well as the file systems carried by the operating systems. The snapshot agent can guarantee that the users can gain data consistency and completeness when they utilize mainstream systems.

OS protection and recovery (DiskSafe): By optionally configuring the client software DiskSafe with the IX3620 series products, the operating system or data on the Windows host can be replicated in real time to the IX3620 device. By so doing, the system achieves continuous protection of the operating system and data. If the Windows operating system is damaged or the host hard disk is faulty, the host can be rebooted with the recovery disks provided with the system. The data can be quickly recovered by connecting to the IX3620 series. If the Windows host of the user is configured with the iSCSI HBA card, the Windows host can be started remotely from the IX3620 device, which ensures fast recovery and continuous operation of the services.

The powerful data management functions of the IX3620 provide unlimited possibilities to the range of diverse applications of users. Regardless of the types of data scheduling, data distribution and data protection needed by users, they can be easily implemented on the IX3620 system.