Case study · Resilience
High-availability IT despite power and network outages
A company with several interconnected sites kept suffering from power cuts and internet outages in the surrounding area. The weak point was not a single server but the connection between the sites. A backup alone would not have solved that.
The situation
Not one fault, but several that could strike at once
The client operates several geographically separate sites that have to communicate reliably with each other for day-to-day business. However, the existing infrastructure was repeatedly hit by disruptions that originated outside the company: power interruptions, internet outages and network faults in the surrounding area.
If a power supply, a line or a central server failed, a site could temporarily lose reliable access to central IT resources. So the real problem was not a faulty server but a chain of dependencies in which any single link was enough on its own to disrupt operations.
A conventional backup alone would not have solved this. A backup protects data – but it does not let people carry on working immediately after a power, line or server failure.
The core of the solution
A separate answer for each failure scenario
The protection covers not one cause of failure but five – each with its own measure that works independently of the others.
Approach
Map the dependencies first, then duplicate selectively
INFONET first analysed the critical dependencies between the sites: which systems does each site need to reach so that people there can work? Only then was redundancy built into the infrastructure – not everywhere, but where a failure actually brings operations to a halt.
Redundant internet connections. Additional backup lines provide an alternative communication path if the primary connection fails. The loss of a single line therefore no longer necessarily cuts a site off completely.
UPS on the critical components. During brief power interruptions, servers, network equipment and central systems keep running. If the outage lasts longer, the systems shut down in a controlled way – which prevents corrupted file systems, interrupted write operations and database problems, the very knock-on damage that causes most of the work after a power cut.
Server synchronisation across two sites. Business-critical server systems are continuously replicated between two sites. Important systems and data are therefore no longer held in only one physical location.
Standby servers. A backup has to be restored first. A prepared standby system is already there and, under the defined emergency plan, can be brought into service much faster. This protects not only the data but also the availability of the services.
Veeam Backup & Replication. Veeam was added as a further layer for backing up and restoring the server systems. The combination of backup, replication, standby system and site redundancy provides far broader protection than a local backup.
Network and firewall
The redundancy had to reach the network too
Two lines are no use if the switchover has not been thought through. That is why the following were taken into account:
- Primary and backup internet line with a defined switchover
- Firewall rule set that applies to both routes
- Site-to-site connections and their routing
- Failover, so the switchover does not have to be done by hand
- Server communication between the sites
- Monitoring of the central components
After implementation
INFONET continues to look after the environment
A high-availability concept that nobody checks is, two years later, a concept on paper only. Depending on the system, ongoing support includes:
- Server monitoring
- Checking the backup jobs
- Checking replication
- Standby servers
- Network and internet lines
- Firewall
- UPS systems
- Fault analysis and recovery when something goes wrong
- Updates and maintenance
Result
A failure-prone infrastructure became one with site redundancy
Dependence on individual power, network and server systems has been reduced considerably. Where a single disruption in the area used to be able to cut a site off from the central systems, a prepared fallback now takes over at each of these points.
The result is a multi-layered business continuity and disaster recovery concept that treats power supply, internet connectivity, server systems, replication and backup together – rather than as four separate topics.
- Redundant internet lines instead of a single connection per site
- UPS protection against power cuts and their knock-on damage
- Server synchronisation between two sites
- Standby servers at both sites
- Veeam Backup & Replication as a separate recovery layer
- Central monitoring and ongoing support from INFONET
Project at a glance
Several sites · redundant internet lines · UPS · server synchronisation · standby servers at two sites · Veeam Backup & Replication · site redundancy · failover · disaster recovery · business continuity · ongoing managed IT services
Frequently asked questions
What prospective clients usually ask
Isn't a backup enough?
A backup protects the data, not your ability to work. After a server failure it first has to be restored – depending on the amount of data, that takes hours. A synchronised standby server, by contrast, is already in place. Only the two together make a robust concept: replication for availability, backup for when data has been corrupted or encrypted.
Do we really need two sites for this?
For genuine site redundancy, yes – otherwise a fire, water damage or a longer power cut hits the production system and the standby system at the same time. If you only have one site, a data centre can serve as the second location. We look at each case to see what is economically justifiable.
What does this kind of protection cost?
That depends on how many systems really need to be highly available. That is exactly the first step: mapping the dependencies and distinguishing between what stops the business immediately and what can wait half a day. Making everything redundant is expensive and almost never necessary. We only give you a reliable figure after this assessment.
How often do you check that the protection still works?
Backup jobs and replication are monitored continuously; the checks are part of our support. Without them, a high-availability concept is no more than an assumption two years later.
A note on confidentiality
For reasons of confidentiality, we only publish client names and contact persons with their express consent. If you have a specific interest, we will be happy to arrange a personal contact after consulting the reference client.
Free initial consultation
How many points of failure does your IT have?
We map the dependencies between your sites and systems and tell you where a single fault would bring your business to a halt today – and which of those points can be protected at reasonable cost.
+49 221 984300-0Switchboard and support hotline
[email protected]Reply within 4 hours on working days
Robert-Perthel-Straße 7250739 Köln – Bilderstöckchen
Mon–Fri 9 am–6 pmEmergency support outside these hours by arrangement