Challenges of service continuity and data integrity

With the rapid development of information technology (IT), information systems play a more and more important role in managing critical services of various industries. Enterprises are searching for a solution on how to cost-effectively ensure the continuity of service systems.

Data is an important asset for enterprises, and storage systems are carriers of data. Therefore, the reliability and availability of storage systems become a vital factor for storage solutions. Storage systems are the core of an entire data center and a single failure in the systems can result in service interruption and key data loss.

Service continuity requirements present three features:

High availability

High availability means that storage systems, networks, and servers coexist in the local infrastructure system and users can access applications when a failure occurs in storage systems, networks, or servers. Therefore, servers, networks, and storage systems must have redundant architecture and robust error tolerance capabilities.

Continuous operations

Continuous operations refer to the non-stop running of services in a failure-free system. Users do not need to stop applications during backup or maintenance tasks.

Disaster recovery

When disasters cause damage to a production center, data and upper-layer applications can be recovered at a remote location.

System architectures protecting service continuity

Based on Oracle RAC application system and the said three features, Huawei launches both local and remote disaster recovery solutions to achieve service continuity. Every solution gives consideration to continuous operations and protects data effectively.

According to different users' application scenarios, Huawei integrates its products and Oracle databases to provide:

• Local high-availability solution on Oracle platforms.
• Point-to-point disaster recovery solution on Oracle platforms.

Local high-availability solution on Oracle platforms:

Local high-availability solution on Oracle platforms adopts hosts, links, and a fully redundant network of storage arrays to achieve the redundancy of systems' key components. Single points of failure do not cause data loss or upper-layer application interruption. This ensures high availability of upper-layer applications. The architecture design maximizes the utilization of storage arrays and high availability of Oracle databases, optimizes the interoperations between storage arrays and Oracle databases, and improves the availability of integrated systems.

The solution provides three redundant architecture options to achieve mirrors between storage arrays:

• Architecture 1: application-layer mirrors. Oracle ASM volume management function is used to achieve mirrors. Main application scenario: Application-layer mirrors provide Oracle database applications and meet the high requirements on RTO and RPO for those applications.
• Architecture 2: network-layer mirrors. Huawei storage virtualization gateway VIS is used to achieve mirrors. Main application scenario: Network-layer mirrors enable the deployment of heterogeneous storage arrays and meet the high requirements on RTO and RPO for service systems to process both Oracle database applications and other applications.
• Architecture 3: array-layer mirrors. HyperMirror/S technology of Huawei T series storage systems is used to achieve mirrors. Main application scenario: Array-layer mirrors satisfy customers' needs for constructing reliable systems at affordable costs and meet the high requirements on RTO and RPO for service systems to process both Oracle database applications and other applications.

Dual-array mirror architecture 1: application-layer mirrors:

Application-layer mirrors use ASM volume management of Oracle database applications to simultaneously store mirror data onto two storage arrays. When one array is faulty, services are automatically switched to the other array. The switchover causes no data loss and no service interruption.

Dual-array mirror architecture 2: network-layer mirrors:

Network layer mirror needs a two-node VIS device. The storage resources of arrays are mapped to VIS and then to hosts. I/O requests of hosts reach to VIS and VIS forwards the requests to two arrays to achieve the real-time array mirror function. When one array is faulty, services are automatically switched to the other array. The switchover causes no data loss and no service interruption.

Dual-array mirror architecture 3: array-layer mirrors:

HyperMirror/S technology of Huawei T series storage systems is used to achieve mirrors. In the architecture, the active array is visible and the standby array is invisible to hosts. Hosts send I/Os to the active array and the active array uses HyperMirror/S to synchronize data simultaneously to standby array. When the active array is faulty, active and standby array are switched manually. The switchover causes no data loss and no service interruption.

Point-to-point disaster recovery solution on Oracle platforms:

The solution uses Oracle Data Guard and HyperMirror/A array technologies to construct the disaster recovery architecture. When a production center is faulty, service systems are quickly switched to the disaster recovery center to protect the continuity of upper-layer applications. The architecture design maximizes the utilization of storage arrays and high availability of Oracle databases, optimizes the interoperations between storage arrays and Oracle databases, and improves the disaster recovery capability of integrated systems.

The solution provides two ways to achieve point-to-point remote disaster recovery on Oracle platforms:

• Data Guard+HyperMirror/A disaster recovery: It is applicable to systems that pose high requirements on RTO and RPO for Oracle databases and require disaster recovery for unstructured data.
• HyperMirror/A of Huawei T series storage systems disaster recovery: It is applicable to service systems that pose high requirements on RTO and RPO to process both Oracle database applications and other applications.；

Oracle Data Guard+HyperMirror/A disaster recovery:

Oracle Data Guard provides disaster recovery to structured data on the application layer. It synchronizes system logs from the production center to the disaster recovery center, and enables the disaster recovery center to provide database applications and protect upper-layer application continuity during production center downtimes. Meanwhile, the importance of unstructured data in service systems requires that the solution integrate HyperMirror/A technology of Huawei T series storage systems to achieve remote disaster recovery of unstructured data. When production centers are faulty, only a little data will be lost, and the disaster recovery center can quickly take over the services.

T series HyperMirror/A disaster recovery solution:

The solution adopts UltraAPM and HyperMirror/A to achieve remote disaster recovery. By using the Oracle data backup and recovery mechanism, the solution categories Oracle database files into indispensable and optional files. The link data compression technology of Huawei T series storage systems satisfies requirements for service continuity and lowers the occupation of network bandwidths. UltraAPM conducts disaster recovery tests and switchovers conveniently and quickly.

Minimized breakdown

• Huawei T series storage systems integrate Oracle RAC technology to provide a local high-availability architecture of full redundancy and high-availability services for 24/7 hours.
• When single points of failure occur in a storage system, services are automatically switched to the other storage system. The switchover causes no data loss and no service interruption.
• When failures occur in a production center, the production center switches to disaster recovery center quickly to reduce service interruptions.

Optimized system performance

• The solution integrates application features of Oracle databases to optimize storage configuration and improve the performance of Oracle database applications running in Huawei T series storage systems.
• RAC technology of databases ensures load balancing among multiple nodes. The cluster constructed by multiple servers of low cost provides high-performance database services.

Effective resource utilization

• The solution adopts Huawei storage products to provide energy-saving technologies such as intelligent CPU frequency control, delicate fan speed control, and intelligent disk spin-down.
• Disaster recovery replication supports link data compression technology to reduce the amount of data transferred on networks.
• The snapshot technology of Huawei T series storage systems conducts tests, backup, and deployment in the disaster recovery center to share the pressure of the production center.

Application-based disaster recovery management

• Application-layer disaster recovery integrates array-layer disaster recovery technologies to ensure comprehensive application-based data protection.
• The application awareness function of Huawei UltraAPM cooperates with storage arrays to process applications and arrays and provides users application-centric management.

Fast deployment and switchover of environments

• HyperMirror/A supports Fibre Channel (FC) networks with a fast transfer rate. It is used for initial synchronization of Data Guard and conducts environment deployment quickly.
• UltraAPM achieves failover within a few minutes to ensure service continuity.

Customization

• Users can adopt different disaster recovery methods according to their needs.

• Minimized RPO and RTO ensure the continuity and reliability of enterprises' services.
• Minimum requirements for disaster recovery bandwidth lower the network construction cost.
• Effective resource utilization maximizes enterprises' benefits.
• Fast deployment and automatic management of strategies stimulate enterprises' innovation.
• The integration with IT systems enables sustainable development of enterprises' IT systems and enhances enterprises' competitiveness.
• The energy-saving technology of Huawei T series storage systems greatly improves power usage effectiveness (PUE), energy saving, and social benefits.
• Storage configurations are optimized for Oracle database applications and the overall system performance is enhanced.