Products
  • OverView
  • Specifications

MacroSAN MS7000 is an member of its new-generation high-end storage product family designed for cloud computing. Underpinned by an advanced design concept and technical architecture, it realizes high performance, safety and reliability and flexible deployment of resources for large-sized data centers.

In addition, it adopts the latest innovative chip technology to build a high-performance and reliable hardware architecture, thus providing a 100G bandwidth with hardware platform enjoying large cache and and strong processing capability for massive concurrent applications.In terms of resource management, it adopts the Cell-based virtualization technology and can flexibly allocate resources through an innovative ICMT management system.In terms of reliability, it adopts the full-redundant dual-control modular design, provides various data protection features and employs the innovative CRAID technology and the IDDC disk diagnosis safety mechanism to reduce down time caused by disk failures and the failure rate by 80%.Besides, based on the Open Data Storage Platform (ODSP), MS7000 can seamlessly integrate the storage system and the user business system through open interfaces, customized specifications and functions, and developing and transplanting industry characteristics.

 

MS7000 view

Highlights
(I) High-performance architecture

In terms of architecture design, MS7000 incorporates cutting-edge technologies, including the dedicated Intel storage processor, ultra-large cache expansion, PCI-E 3.0, SAS 2.0, and SSD to build an end-to-end high-performance storage platform.

MS7000 integrates a dedicated Intel storage processor to boost the processing performance. The processor integrates the XOR engine to realize RAID hardware acceleration, which not only significantly improves the RAID's processing efficiency, but also reduces the workload of the processor.In addition, it also integrates a memory controller to enable direct communication between the processor and the memory without the Northbridge connection, which greatly improves the access speed and expands the communication bandwidth to more than 256Gb/s.

In the internal controller, MS7000 adopts the PCI-E 3.0 channel technology.The bandwidth of a PCI-E3.0 channel is 8.0Gb/s. Multiple PCI-E3.0 channels expand the host channel bandwidth, the cache mirror channel bandwidth and the end disk channel bandwidth to over 100Gbps, thus ensuring end-to-end high performance.

High-performance SSD drives have won wide applauses among the industry. MS7000 integrates the cutting-edge disk technology into its high-performance architecture to boost the IO response capability of the drives.In addition, the HotCache and the automatic hierarchical storage system can perfectly blend the data hotspot with the SSD drives to give a strong boost to the overall performance only with a few SSD drives.

Caching is the core system of a storage array and also one of the most important factors that affect the performance of a storage controller.Due to technical limitations, the conventional storage array has a small cache capacity and poor scalability with the complicated tuning management.In the cloud computing environment, the stored concurrent I/O access volume grows rapidly, the data access hotspot changes frequently and the demand for storage cache capacity increases dramatically.Meanwhile, the large-capacity cache requires more efficient storage cache management.

MS7000 supports ultra-large-capacity cache and employs a leading intelligent cache management system to greatly improve the caching efficiency.

Capacity configuration as required: The cache is a major indicator for storage arrays. By increasing the cache, users can significantly improve the storage performance.MS7000 can flexibly configure the size of a cache according to users’ needs.This is more cost-effective than buying better but more expensive products.

Global intelligent caching: MS7000 incorporates the intelligent caching system to manage the global cache.Read and write caches of a controller can be dynamically adjusted and the caches of two controllers cannot be allocated symmetrically to give full play to the performance of controllers.In terms of specific caching policies, the read and write caches are divided into reservation, sharing and burst blocks to meet requirements for real-time performance changes to LUNs.In this system, LUN can obtain a large cache under great pressure or a small cache space under small pressure, thus eliminating performance bottleneck for the LUN and maximizing the utilization of the cache of the entire system.

Safe and reliable data: MS7000 uses the mirror system and provides the power-off protection function to ensure the safety of cache data.In case of an external power failure, data in the write cache can be stored in the first four drives (reservation drives) of the storage array. When the power supply is retored, data can be recovered from the storage disk.

(II) Cell-based intelligent resource management

Cell, an active data unit, is a basic unit for the MS7000 to manage storage resources. Using the "cell" mechanism, the MS7000 changes the traditional LUN management mechanism at the bottom layer of data blocks and builds a complete virtual storage architecture to support delicacy management and flexible scheduling of resources.

Traditional logical storage resources are managed based on LUN levels. There is a vertical relationship among LUNs, RAIDs and drives. Once a LUN is allocated, the space of the LUN corresponds to the fixed RAID group and drives. In this case, it is hard to improve the performance, expand capacity of the LUN, or change its configuration. Any change to resource configuration leads to a lot of changes in other parts. Even worse, the entire application may need to be recreated.

Intelligent cell management technology (ICMT) is an innovative technology developed by MacroSAN. With ICMT, the LUN does not correspond to a fixed RAID group, but the logical space with multiple cells. The LUN does not care about the RAID type or the actual locations of the cells. The MS7000 manages logical resources based on cells. This completely eliminates the binding relationship among LUNs, RAIDs and drives. The cell technology forms a virtual architecture in a storage array.

Flexible capacity configuration: During initial capacity allocation, the system automatically obtains storage space from the cell pool without the interaction of administrators. Then, the system enables automatic increase in space by the thin provisioning technology. For example, the application system requires 1 TB logical space, while the storage array allocates only 500 GB physical space. When the actual capacity of applications approximates to or reaches 500 GB, the storage array assigns new physical space, for instance, 200 GB, to the application system according to the pre-defined policy. Thus, the actual physical space reaches 700 GB. The ICMT and thin provisioning technology enable the application system to raise a request for space to the storage array regardless of space restrictions and physical locations of the resources. The storage system automatically and flexibly configures and manages capacities.

Automatic storage tiering: ICMT eliminates the binding relationship among LUNs, RAIDs and disks. Through cell-based data copy and migration, ICMT enables flexible data transfer among different disk mediums according to data access frequency to support automatic storage tiering. For example, hot spot data is stored on SSD disks, non-hotspot online data on high-speed SAS disks, and archived/backup data on low-speed SAS disks, and archived/backup data is set to sleep mode on low-speed SAS disks. ICMT and automatic storage tiering policy thoroughly eliminated the performance bottlenecks brought about by dynamic hotspot data.

(III) Innovative CRAID technology

The traditional RAID operates based on disks. If a few blocks of a disk are damaged, the entire disk is kicked off from the RAID group and data on this disk is reconstructed on the hotspare disk. As disk capacity increases greatly, such a RAID mechanism is challenged because data reconstruction requires a long time and consumes a lot of system resources. For example, reconstructing a 2 TB the RAID 5 SATA disk requires 20 hours when there is no IO to disks, and requires several days or more than a week when there is IO to disks.

Cell-based RAID (CRAID) is an innovative technology developed by MacroSAN. With CRAID, after a traditional RAID group is created, the system divides the RAID space into small data blocks, called cells, and then manages resources based on cells. For example, if some data blocks of a disk are damaged, only the cell that contains the damaged data blocks is affected. The system reconstructs data of this cell only without affecting other data on the disk. The CRAID enables delicacy management of disks based on cells. This enhances tolerance of hard disk faults and significantly raises data reconstruction efficiency. Details are as follows:

Tolerance of faults of multiple disks in one RAID group: In a RAID group, as long as two data blocks in one cell do not fail at the same time, the RAID group allows medium damage of multiple disks without data loss.

Reconstructing cells that contain data only: If data reconstruction is required for recovery from a serious disk fault (such as a physical fault), the CRAID only reconstructs the cells that contain data. The idle cells are not reconstructed.

Local reconstruction: In the case of a temporary failure (for example, a disk is removed and then inserted), the CRAID only reconstructs the cells with data changes instead of the entire hard disk.

Rapid reconstruction: In the case of incomplete damages of a disk (such as, errors of some mediums), the CRAID only checks and reconstructs the cells with medium errors. Other data is reconstructed by directly copying from the original disk.

(IV) Comprehensive disk security policies

Traditionally, storage security means system redundancy and data protection, but neglects disk security. According to statistics, over 90% hardware faults of the storage system are caused by disk faults. In this way, disk security is the essential for storage security.

The MS7000 has a comprehensive disk security mechanism. With its IDDC, the MS7000 initiatively prevents disk faults, rapidly repairs faulty disks, and reduces incorrect judgment of disk faults through fault recovery and troubleshooting, reducing breakdown time caused by hard disk faults by 80%.

Initiative disk detection: The traditional disk recovery is to correcd t a fault after it occurred. On the contrary, the disk diagnosis mechanism of the MS7000 is to detect faults prior to fault occurrence. During normal system running, the disk detection module regularly detects RAID check information of data disks and scans the entire blank disk and hotspare disk. As a result, the disk detection module can predict disk faults earlier than the application system and start corresponding activity in advance.

Rapid fault recovery: The MS7000 provides multiple data recovery policies for a detected disk fault. For pseudo disk faults, the MS7000 can recover faults by automatic power-on/power-off of disks. For a few damaged blocks of disks, the MS7000 can use Remap and RAID/CRAID mechanism to replace them. For a lot of damaged blocks of disks, the MS7000 can rapidly rebuild them by CRAID.

Recheck of disk faults: Incorrect judgment of disk faults is common to the storage system. To reduce incorrect judgment of disk faults, the disk detection center will recheck the disk to reduce the incorrect disk detection.

(V) Dual controllers for full redundancy and high reliability

Regarding hardware architecture, the MS7000 holds two controllers for full redundancy and adopts a modular design. The MS7000 supports end-to-end redundancy of data channels and major modules to allow for hot swapping and rapid replacement.

Fully-redundant architecture: When data is written to a disk, the data goes through the host channel, controller, cache, disk channel, and disk chassis, and finally reaches the disk. In the entire data path, the MS7000 employs a full-redundant architecture to guarantee system reliability. The MS7000 holds two controllers. When a single controller fails, services are automatically switched over to the other. The MS7000 uses a cache mirroring mechanism. Two copies are maintained for the data in the write cache, and the data is sent to the reserved disk after in case of power-off. The MS7000 supports ALUA multi-path mechanism to allow host path switchover in case of a fault. The MS7000 supports dual SAS links to provide two redundant external ports for each SAS disk chassis. When one SAS link fails, path switchover automatically starts. In addition, each SAS disk has two ports connected to two data channels for redundancy and mutual backup.

Dual active controllers: The MS7000 is equipped with two controllers to support redundancy and load balancing. The two controllers are both active and provide service independently. When a controller encounters a performance bottleneck, part of its services are automatically migrated to another one for load balancing.

Modular design: Major components of the MS7000, such as chassis, controllers, power supply systems, fans, batteries, and host interface cards are modular. Hot-plug and on-line replacement of an independent component is allowed. When a component fails, the fault can be rapidly isolated and the component can be fast replaced to avoid impact on the entire system.

Various host interface cards: Each controller of the MS7000 is configured with multiple host interface card slots. Four kinds of interface card are available: 1 Gbit/s iSCSI, 8 Gbit/s FC, 10 Gbit/s iSCSI, and 10 Gbit/s FCoE. You can flexibly choose host interface cards to easily establish an IP SAN, FC SAN, or hybrid SAN storage environment.

Perfect power supply and cooling systems: The MS7000 is equipped with redundant power supply systems. The power module provides perfect device protection against output over-current, output short circuit, and output over-voltage. Turbofans of the MS7000 can adjust the rotation speed of themselves automatically according to the temperature. Symmetrical upper and lower axes of the MS7000 controller not only enhance cooling effects of fans, but also reduce the impact of a single-fan failure.

(VI) Diversified data protection functions

On the basis of full-redundancy and perfect disk security policy, the MS7000 provides diversified data protection functions, including continuous data snapshot, data replication, and data mirroring. These functions enable online, nearline, local and remote data management and protection, thus delivering hierarchical and cross-region storage solutions.

Snapshot: The continuous data snapshot function of the MS7000 creates incremental copies of data volumes at multiple historical time points. In case of soft errors, such as data damages caused by software programs or viruses, and file deletion by mistake, data can be rapidly recovered by rollback to appropriate time point flags. This function is especially suitable for continuous data protection for critical services.

Replication: This function supports remote data backup. In case of a disaster, the MS7000 can rapidly recover data to guarantee business continuity of customers. The MS7000 can seamlessly interconnect with WAN through IP interfaces for cross-IP-WAN remote disaster recovery without protocol conversion, thus effectively lowering the costs on disaster recovery links.

Mirroring: The data mirroring function of the MS7000 creates data mirroring between two storage systems to produce a completely synchronous real-time mirroring for primary data. Each write I/O request is concurrently saved to the primary storage and secondary storage. When the primary storage fails, the secondary storage will take over services.

(VII) User-friendly management

The MS7000 provides graphical user interface (GUIs) and command line interfaces (CLIs) to facilitate user’s operation. The application-oriented wizard interfaces enable administrators to set application types and capacity by a click, thus easily allocating resources.

The MS7000 supports EventLog, indicator alert, console alert, buzzer alert, SNMP alert, email alert, and SMS alert. Regarding alarm processing, the MS7000 supports custom fault-event-and-alarm linkage. Administrators can set different alarming modes for each type of event.

The MS7000 supports centralized management. It enables status query and alarming through SNMP traps and provides interfaces for calling WebService for major management functions, thus helping third party applications to carry out unified management.

The MS7000 offers a perfect performance monitoring mechanism, which can monitor the performance of controllers, LUNs, cells, RAIDs, and disks, and delivers diversified performance reports.

(VIII) ODSP

The ODSP of the MS7000 seamlessly integrates the storage system with customers’ service systems by open interfaces, custom specifications and functions, and developed and migrated industrial features, adding to the transparency and interconnectivity of the storage system.

The MS7000 employs a modular design based on the latest Intel hardware platform. Its components can be configured on demand and hardware interface cards can be customized. Regarding software architecture, the MS7000 can provide an open application programming interface (API) for embedding functions and features and secondary development. The MS7000 supports GUI-based centralized interface management and offers comprehensive command line interfaces/SNMP/WebService interfaces. With its strong R&D teams, MacroSAN delivers technical support at R&D level and custom development, and can rapidly respond to your need.

(IX) Heterogeneous storage virtualization

MS7000 has a built-in virtualized data management engine, which can store data on different storage devices such as IP SAN and FC SAN in MS7000's unified resource storage pool.By integrating storage arrays of different brands and architectures, users only need to focus on logical storage resources provided by MS7000 rather than back-end storage devices. This can effectively simplify the management, cut maintenance costs, and improve utilization of resources.

(X) MS7000 base on ODSP unified software platform

which is compatible with the full range of MS products, So MS product support for a smooth upgrade. By changing the controller, the other MS products to upgrade to MS7000, without the need for data transfer in the process of upgrading, effectively protect the user's investment.

Item

MS7000

Architecture

Virtualized multi-engine architecture

Number of Controllers

2-16 (optional)

CPU

One-path or two-path Intel multi-core processor for each engine

Max Cache

6144GB

Max Hosts

384

IO Module

4-port 1 Gb/s iSCSI module

4-port 8 Gb/s FC module

2-port 16 Gb/s FC module

2-port 10 Gb/s iSCSI module

2-port 10 Gb/s FCoE module

4-port 4*6 Gb/s SAS module

4-port 4*12 Gb/s SAS module

Drive Interface Type

24/48Gb SAS interface, each interface containing four SAS channels

Drive Types Supported

SSDSASSATA

Max Drives

3600

Max Number of LUNs (Pool)

64000

Disk detection and diagnosis

Supports periodic disk detection

Supports intelligent dynamical adjustment of disk speed detection

RAID Options

RAID/CRAID (CRAID2.0) 0, 1, 10, 5, 6

Supports the dedicated hot standby, global hot standby and idle disk hot standby

CRAID

The CRAID group allows medium error in many disks and physical failure in any three disks, and supports common rebuilding, local rebuilding and rapid rebuilding.

LUN synchronization feature

Supports non-synchronization, check synchronization and rapid synchronization

Compatible operating systems

AIX, HP-UX, Solaris, Windows, Linux, VMware etc.

Host multi-path

Be compatible with the multi-path software supporting ALUA characteristics

Basic management software

MacroSAN management pack, which has such functions as basic storage management, CRAID, system monitoring, logging and alarming

Management interface

Supports a Chinese graphical interface and a command-line interface, and provides a WebService access interface

Snapshot

Supports 512 snapshots in single volume, snapshot rollback and snapshot view

Replication

Supports the IP-based remote duplication, and provides functions of 1:2, hopping, and 64-to-1 duplication

Advanced features

Mirroring, performance monitoring, thin provisioning, intelligent hierarchical storage, automatic data migration, heterogeneous storage virtualization

Power supply

AC100V120V/200V240V60Hz/50Hz

Temperature Gradient

Operating temperature: 0°C - 35°C; 10°C - 35°C (recommended)

Non-operating temperature: -20°C - 60°C

Humidity

Operating humidity: 20%-85% without condensation

Non-operating humidity: 10%-90% without condensation