High availability
Optional active or passive high availability is available for our server solutions. With the help of our solutions for high availability, the availability of the respective services can be maximized.
Available solutions
We differentiate between passive high availability and active high availability in our solutions. All failover solutions are operated by a homogeneous server infrastructure under identical conditions, so that every failover concept can meet the identical hardware and network requirements.
Passive high availability
When passive high availability is activated, the respective V-Server is continuously replicated to another physical server within another physical server rack. We offer replication intervals of 5 minutes, 10 minutes, 15 minutes, 30 minutes, 60 minutes, 90 minutes and 120 minutes. If the primary physical server (hypervisor) fails, a failover is automatically triggered via the quorum procedure and the replicated server is started.
The failover process can take up to 120 seconds. The corresponding servers are then restarted on an alternative server.
A major advantage of passive high availability is that compatibility with the services used is not necessary compared to active high availability. Thanks to the technology used, the respective servers are only available once in our data center so that the actual application does not have to provide any support for high availability.
In addition, the software licenses used do not have to be booked twice, as the connected servers run on either the primary or secondary hypervisor.
As soon as the primary hypervisor is available again, the servers are migrated live so that there is no further downtime due to the "reallocation process".
Active high availability
Active high availability enables a failover process without any downtime by providing so-called hot standby servers. In addition, we provide a TCP load balancer with a floating IP address that is also highly available and can be operated by the secondary server rack in the event of a fault. The support of the underlying applications is required for the realization of active high availability. MySQL databases, for example, must be continuously synchronized with the entire infrastructure.
If there is a fault in the primary infrastructure, the secondary TCP load balancer takes over the routing of the network packets and forwards incoming network requests to the secondary infrastructure.
The applications of both the primary and secondary infrastructure must support this concept of high availability so that synchronization of the data statuses of all servers can be guaranteed at all times. In addition, the application used must support the master delegation procedure in order to be able to recognize the failure of the primary infrastructure.
Advantages & disadvantages
We have summarized the most important advantages and disadvantages of both high availability solutions below. Depending on the use case, both solutions can have further advantages and disadvantages.
**Advantages
| Passive high availability | Active high availability |
|---|---|
| Simple maintenance | No downtime |
| no application support required | high availability up to 99.99 % |
| lower costs (cf. active high availability) | maintenance can be carried out during the day |
**Disadvantages
| Passive high availability | Active high availability |
|---|---|
| Downtimes of up to 120 seconds | High maintenance effort |
| High availability up to a maximum of 99.9% | Application support required |
| Possible data loss between replication | Higher costs (TCP LB + software licenses) |
If you are interested in one of the two solutions, please contact us at contact to discuss the next steps. We will be happy to advise you on choosing the right variant and work with you to develop a concept for your application.
High availability classes
High availability classes of the BSI (Federal Office for Information Security)
| Class | Availability | Max. Downtime per year |
|---|---|---|
| 0 | Without guaranteed availability | - |
| 1 | 99.0% | < 3 days, 15 hours, 40 minutes |
| 2 | 99.9 % | < 8 hours, 46 minutes |
| 3 | 99.99 % | < 53 minutes |
| 4 | 99.999 % | < 6 minutes |
| 5 | Disaster Tolerant | Function must be guaranteed under all circumstances |
AEC classes
Availability Environment Classification (AEC)
| Class | Name | Description |
|---|---|---|
| AEC-0 | Conventional | Function may be interrupted and data integrity is not essential |
| AEC-1 | Highly Reliable | Function may be interrupted, but data integrity must always be guaranteed |
| AEC-2 | High Availability | Function may only be interrupted minimally or within defined times during main operating hours |
| AEC-3 | Fault Resilient | Function may only be interrupted outside of defined operating times / main operating times. |
| AEC-4 | Fault Tolerant | Function must be maintained at all times, 24/7 operation (24 hours, 7 days a week). |
| AEC-5 | Disaster Tolerant | Function must be guaranteed even under unusual circumstances (e.g. in the event of a disaster). |