Active-Active, as the name implies, means that both sides are active online to provide services, as opposed to the traditional Active-Standby model. A true Active-Active solution should cover all levels of infrastructure, middleware, and applications.

Dual data centers simultaneously provide business production services to the outside Active-Active mode, the two data centers are peer-to-peer, regardless of master and slave, and can be deployed at the same time, which can greatly improve the utilization of resources and the efficiency of the system, performance, so that customers can get the maximum value from the disaster recovery system.

a. Two production centers deploy the same business system, combined with the network layer, host layer or application load balancing technology, to achieve parallel work and load sharing of business systems in the two data centers.

b. Two production centers deploy different business systems and take over each other's real-time disaster recovery.

Data center Active-Active is further divided into: same-city Active-Active, off-site Active-Active.

Disadvantages of the traditional primary backup model

For the purpose of disaster recovery, two (or more) data centers are generally built. One is the primary data center to take care of the user's business, and the other is the backup data center to back up the data, configuration, business, etc. of the primary data center.

There are generally three types of backup methods between the primary and backup data centers: hot backup, cold backup, and Active-Active.

 In the case of hot standby, only the primary data center undertakes the user's service, and at this time, the backup data center performs real-time backup of the primary data center, and when the primary data center hangs up, the backup data center can automatically take over the service of the primary data center, and the user's service will not be interrupted, so the switching of data centers will not be felt.

● In the case of cold standby, only the primary data center takes over the business, but the standby data center does not perform real-time backups of the primary data center, which may be performed periodically or not at all, and if the primary data center hangs up, the user's business will be interrupted.

● Active-Active is to think that it is too wasteful for the standby data center to only do backup, so let both the primary and standby data centers take on the user's business at the same time, at this time, the primary and standby data centers are backups for each other and real-time backups are performed. Generally speaking, the load of the primary data center may be more, such as sharing 6070% of the business, and the backup data center only share 40% 30% of the business.

The traditional primary backup mode is that a business runs in only one data center. Enterprises deploy a large number of backup servers in the backup center in conjunction with the disaster recovery level requirements and business requirements, but the backup center only provides disaster recovery services for the business, and only when a disaster occurs and the production data center is down, the business systems in the disaster recovery center start up these servers, which results in the wastage of the server resources in the backup center, and the WAN links cannot be fully utilized. WAN links cannot be fully utilized.

Advantages of Active-Active Data Center

Fully utilizes resources, avoiding the waste of a data center being idle all year round. Through resource integration, the service capacity of “Active-Active” data center is doubled.

Active-Active data centers are imperceptible to users if one data center is down while the other is still running.

In the model of one disaster recovery center, if the production data center is down, it takes half an hour, or even two hours, or even longer to start the disaster recovery center, and in the time it takes to start the disaster recovery center, user transactions will be seriously damaged.

The biggest advantage of Active-Active data center is the effective use of resources. The investment in the construction of the disaster recovery center is huge and the annual operation and maintenance costs are extremely high. If the resources are in an idle state, the resources are quite wasteful, and with virtualization, the idle resources can be integrated, and the service capacity will be doubled. Many resources in the banking system are elastic demand, such as funds, precious metals transactions, electronic payments, and Internet banking transactions, in the transaction of the day when the transaction volume may reach the total annual transaction volume. Therefore, when planning the capacity of the banking system, the peak transaction is fully taken into account, but there is a great deal of waste of transactions in normal time. Take Taobao's “Double Eleven” campaign as an example, the transaction volume may reach the sum of the annual transaction volume in a few minutes, which requires the system to increase its service capacity by ten times, and Active-Active is the best choice for this. This is where Active-Active data centers and flexible and fast resource scheduling come into full play. Cloud computing technology, so that the IT system has the ability to integrate resources, so that the system has full elasticity, at any time you can schedule ten machines to improve the service capacity to ensure that the transaction of the sudden demand, as well as a variety of unexpected factors caused by the transaction volume surge.

With cloud computing technology, does not mean that the investment will be less, but the utilization of resources will be higher, the system but the shock resistance will be stronger, the free scheduling ability will be stronger.

Automation is “Active-Active” and “cloud computing” is an essential prerequisite

Cloud computing requires automation to help system maintainers automate resource provisioning. For example, through the virtualization technology virtualized tens of thousands of virtual machines, the daytime need 50 machines to the Internet banking system to provide web services, the night of the Internet banking transactions less, more precious metals transactions, the 50 machines to be deployed to another system. This 50 units can not be a person one by one deployment, that may be a night with the allocation can not be completed, the need for automated software to automatically adjust the allocation of resources.

Difficulty of “Active-Active” in different locations

Of course, the deployment of “Active-Active” data center is also very difficult, especially off-site “Active-Active”, involving data synchronization efficiency. If the data synchronization efficiency is not up to the requirements, in the event of a disaster, it will cause a period of transaction loss. In the off-site “Active-Active” model, both data centers accept transactions at the same time, which is technically difficult and requires numerous changes to the underlying procedures.

Three conditions for the construction of Active-Active data center The construction of Active-Active data center must first meet three conditions, the first one is application Active-Active, that is to say, the database must be Active-Active, the second one is network Active-Active, and the business network must ensure that it can connect to two data centers at the same time, and the third one is that the business network must ensure that it can connect to two data centers at the same time. The second is that the network should be Active-Active, the business network should be able to connect two data centers at the same time, and the third is that the data should be Active-Active, the data on both sides should be able to be used independently.

Disadvantages of Active-Active Data Center Solution

Although Active-Active disaster recovery solution for centrally managed data centers ensures the linearity of business production to a greater extent and effectively defends the ability to recover business production from catastrophic events. However, Active-Active data center disaster recovery solution still has certain shortcomings, the ideal and the reality there is always a certain distance.

1. Fracture phenomenon

Active-Active data center solutions to achieve site-level redundancy disaster recovery solutions, but limited by current technology and other factors, in the construction process to solve the business continuity problems currently faced by enterprises, but also generated new problems, is the Active-Active solution prevailing cracked brain phenomenon, in the event of an accident, if the monitoring technology is not in place, the system platform is not healthy, and the two sides are not able to recover business production. In the event of an accident, if the monitoring technology is not in place, the system platform is not healthy, the two data in the network volatility interruptions and other factors occur, so that the two data centers integrated business systems will be split into two independent data centers. This makes it difficult for users to decide which is the only production data and which is the non-production data that will be scrapped. This is the early years of veritas VVR solution out of the disaster recovery stage is one of the reasons.

2. Non-“zero loss”, no soft error protection

Active-Active disaster recovery solutions emphasize the advantages of a healthy operating platform, large-scale disaster occurs is “zero” data loss, but if the Active-Active platform itself is not healthy or suffered a logical failure, and can not guarantee zero data loss. Data recovery from such a failure or gradual disaster occurs, but also need to use the backup system data recovery means or methods. Therefore, Active-Active disaster recovery program in most cases does not have the guarantee to solve the soft error, but precisely the probability of such an event occurs far more than a site-level disaster and hardware failure events. In 2012, a provincial government department's business system has been built disaster recovery system, but in the business system for upgrading errors, resulting in business downtime for more than a week, and most of the time during this period is to find the basis for the recovery of data.

3. Need to tolerate high reliability and performance degradation

Although the Active-Active disaster recovery solution improves site-level redundancy protection, in practice it does in addition to lower the reliability and performance of the overall business platform. In the reliability program, Active-Active disaster recovery solution is the local dual-cabinet hardware redundancy program across the site construction, whether it is a traditional cluster system, virtualized host platform Vmware, or Oracle RAC, etc., cross-site construction will invariably add a few points in the business platform in the unstable factors, I think from the now popular all-in-one solution! More illustrative of this aspect of the problem, that is, the simpler the more stable the system. In the performance of the program, the number of network delays between the site monitoring, synchronization of business sessions to confirm the number of network delays, coupled with data synchronization double writing fiber optic delays, are more or less affected by the overall performance of business processing. The farther away the impact is more obvious, if the distance is closer, will also lose the significance of building Active-Active disaster recovery data center.

4. Operation and maintenance is not simple

Active-Active disaster recovery solution disaster switching becomes simpler, but in the actual maintenance is not simple, in addition to the requirements of enterprise users to enhance their own maintenance capabilities, but also need Active-Active disaster recovery solution provider's after-sales service capabilities.

a. Enterprises must strengthen their own personnel maintenance capabilities, in order to have the ability to maintain cross-site Active-Active system, that is, the need for business users to maintain their own personnel must be changed from the ability to maintain equipment to have the ability to maintain the Active-Active system architecture, in order to stabilize the normal operation of the system, so that the Active-Active system to achieve the effect of that The effect of the Active-Active system can be realized.

b. The service capability of the provider also has a direct impact on the effect of Active-Active disaster recovery system deployment, in the existing cases, we often see the 800 phone number of the provider, in addition to the collection of logs or the collection of logs, in addition to the background diagnostics or background diagnostics, often a small problem requires a good number of layers, times of communication in order to solve the problem, how to ensure that the Active-Active disaster recovery system in such a way? How to guarantee the stability of Active-Active disaster recovery system in such a way? How to ensure the stability of the Active-Active disaster recovery system in such a way? How to ensure that the user's expectations of the Active-Active system in the linear requirements?

5. Cost-effective and not too high

We often hear that Active-Active disaster recovery solutions can make production centers and disaster recovery centers are “live” up, the effective use of resources, in the face of catastrophic events, maximize the linearity of the business system, lifting the original disaster recovery system is not prepared for a disaster, and so on the shortcomings. However, when we seriously consider the construction of Active-Active Disaster Recovery System, we found that if the maintenance capability of our own IT staff is insufficient, it is difficult to achieve the desired results. In reality, many users one-time cost to build the system, the subsequent maintenance funds are difficult to apply for, this situation is difficult to effectively protect the healthy operation of our information systems. The Bank of Ningxia is in the case of no follow-up maintenance funds support, hardware failure, their own IT staff to repair the process of human error and caused major accidents. Therefore, the construction of Active-Active disaster recovery system at the same time, it is necessary to protect the subsequent maintenance funds. Make Active-Active disaster recovery system to the high level of bias.

General domestic IDC, financial services will be considered two places, three centers, non-financial services will be considered as the main backup IDC or Active-Active IDC, if the global business, the general will be divided into regions around the world to establish a global IDC, and a regional at least 2 nodes to do the main backup or Active-Active.

For network products list and quote, please visit: https://www.hi-network.com/categories/cisco or contact us at www.hi-network.com  (Email: [email protected] (en inglés))