---
canonical: "https://safekit.eviden.com/wp-content/uploads/documents/wp-hadecision-en.pdf"
llms_index: "https://safekit.eviden.com/llms.txt"
llms_section: "Datasheets & White Papers"
topics: "High availability decision white paper"
---

# PDF Converted to Markdown

## Page 1

```
High Availability (HA)
Choosing a Solution
```

## Page 2

```
02 - High Availability (HA) Choosing a Solution
Summary0
3 Case study: a crisis in an airport
05     More information on our web site
05 3 partner testimonials
06 10 reasons to choose the SafeKit clustering software
What to look for when choosing a 
high-availability solution?
This white paper is based on a case 
study: a crisis in an airport.
02 - High Availability (HA) Choosing a Solution
```

## Page 3

```
Case study: a crisis in an airport
Every computer-system-based activity, 
both big and small, is one day or the other 
faced with the problem of computer failure. 
Unfortunately, the day the failure occurs, 
a small problem may turn into a general 
crisis due to a succession of errors. This 
was exactly what happened at an airport, 
despite the fact that it was equipped with 
a redundant solution. This story will enable 
us to highlight the main characteristics of a 
good-quality high-availability solution.
The crisis occurred at an airport on the 
traveler-management application. This 
application displays information on the 
airport’s notice boards, making it possible 
for travelers to locate their boarding gate or 
to know where to retrieve their luggage.
High availability based on 
redundancy
The size of the traveler-management 
database is not big. But this database is 
extremely sensitive and must withstand the 
worst case scenario at an airport: aircraft 
crash-landing on a computer room.  
The selected redundancy solution consisted 
of two separate servers located in two 
remote computer rooms. Each server had 
a local traveler-management database, 
and a replication tool was used to ensure 
database redundancy between the main 
server and the backup server.
The incident
To maintain the database with the new 
flights for one week, every Sunday the 
database is updated with the flight 
schedules for the following week. 
Unfortunately, one Sunday, while the flight 
update operation was in progress on the 
main server, replication stopped on the 
backup server. So, at the end of the update 
operation, the main server contained the 
flights for the following week, and the 
backup server the flights for the previous 
week. 
On Monday, the main server supplied the 
right flight schedules to the airport’s notice 
boards. For the IT department, airport 
management was working properly, despite 
the undetected absence of replication on 
the backup server.
The crisis
In the middle of the week, the main server 
rebooted automatically as a result of an 
error. Unfortunately, it remained blocked 
in its booting process and failed to reboot 
the traveler-management application. The 
notice boards at the airport went blank, and 
the IT department was alerted. 
Since the main server was blocked in its 
booting process, a decision was taken to 
restart the application on the backup server. 
The application was then restarted on the 
backup server without any difficulties and, 
a few minutes later, all the notice boards at 
the airport became active again. But they 
displayed to travelers the flight schedules 
for the previous week! 
The red alert at the IT department soon 
turned into a black alert. The traveler-
management application was immediately 
stopped on the backup server, and the IT 
team decided to solve the problem of main 
server reboot. Meanwhile, the airport’s 
information-management application 
remained unavailable to all travelers. 
The situation gets worse
A few hours later, the administrators 
succeeded in solving the problem on the 
main server, and in rebooting this latter. 
But, unfortunately, in the booting phase, the 
replication tool on the main server started 
automatically. It detected that replication 
was already active on the backup server 
and resynchronized the main server’s local 
database from the backup server. The 
database of both servers thus contained 
the flight schedules for the previous week!
Airport activities were disrupted for a 
whole day, and the main server could only 
be correctly resynchronized through the 
special Sunday operation. 
In the end, the airport replaced the 
replication tool with a complete replication 
and high-availability solution, based on the 
criteria described hereinafter. 
Where are the vulnerabilities of 
a high availability solution?
The problem encountered by the airport’s 
IT department was due to the replication 
solution chosen. Replication was working, 
but the solution was incomplete in terms 
of high-availability and recovery in case of 
failure. Yet in its technical data sheet, this 
solution is said to have high-availability 
mechanisms with heartbeat and recovery 
scripts! 
Therefore, it is necessary to properly 
identify the high-availability needs while 
choosing a replication solution. The needs 
expressed by the production team after this 
crisis at the airport are as follows. 
Every administrator must know 
that replication has stopped on 
the backup server
The first event that led to the crisis at the 
airport was the updating of flight schedules 
on the main server on Sunday while the 
backup server was out of service. 
On Monday, when the entire production 
team arrives, a simple glance at the 
replication application’s administration 
console should be enough to detect the 
absence of replication on the backup server. 
Therefore, the replication solution must 
have an administration console which can 
remotely connect to servers and provide a 
summary of the status of a high-availability 
application on two servers. It must equally 
be possible to send mail and to integrate 
the product in the administration console 
used by the customer if replication stops. 
It should be very easy for every 
administrator to restart 
replication on the backup 
server
On Monday, upon detection of replication 
stop on the backup server, any 
administrator from the IT department 
should be able to restart the replication: 
• either by clicking a button on the 
replication administration console, or
• through a very simple online command 
offered by the replication solution on the 
backup server, or
• by rebooting the backup server. 
Restart should never be reserved to a 
replication solution expert. In fact, the 
system must be made highly available 
again even when the expert is not present. 
High Availability (HA) Choosing a Solution - 03
```

## Page 4

```
It should be possible to 
resynchronize the backup 
server while the application is 
running on the main server
On Monday, when the IT department 
resumes work, the passenger-management 
application at the airport is running on 
the main server, and replication must be 
restarted as quickly as possible on the 
backup server. It should, therefore, be 
possible to resynchronize the backup-
server database once the main server is 
accessed. 
Popular file replication solutions cannot 
resynchronize a backup server without 
stopping the application on the main 
server! These products implement 
replication solutions but absolutely not 
high-availability solutions.
If a server’s local databases are 
not updated, the solution must, 
by default, refuse to run the 
application on the non-updated 
server
In the case of the airport, the alert turned 
into a crisis because this feature did not 
exist. 
The backup server with the flight schedule 
for the previous week was not up-to-date, 
still nothing stopped the application from 
starting on this server. 
Later, when the main server was rebooted, 
data synchronization took place in the 
wrong direction in that both servers were 
synchronized with the flights for the 
previous week, thus resulting directly in the 
crisis at the airport. 
Recovery-control mechanisms must enable 
the administrator to avoid human error, 
which a simple replication solution does not 
offer. 
In case of failure, replication 
should not result in the loss of 
data about already registered 
travelers
This important requirement is not met with 
the asynchronous replication traditionally 
implemented by replication solutions! Take 
care and check that the replication solution 
is synchronous.
In fact, in case of failure with an 
asynchronous replication solution, you 
have to locate the passengers registered 
for a flight prior to the failure, but whose 
details are not saved in the backup server 
database due to asynchronous replication. 
These passengers’ reservation is lost after 
the failure, and their seats are again free in 
the reservation system! 
Data replication on the backup 
server must be synchronous to 
ensure high availability
With a synchronous replication solution, 
data is not lost in case of server failure. In 
fact, at the check-in counter, a passenger 
is checked in by reserving a seat for him 
or her on the aircraft. The reservation 
request is sent to the main server, and 
the passenger-management application 
permanently stores the reservation-related 
information on the local disk. 
In synchronous replication mode, when a 
piece of information is permanently stored 
on a server's local database, it is also stored 
on the remote database.
Thus, when the check-in counter receives 
the acknowledgement of reservation and 
the attendant releases the passenger by 
issuing him or her a boarding pass, the 
reservation is permanently recorded in both 
servers’ databases (this is not the case with 
asynchronous replication).
If the main server fails, the passengers 
currently being checked in are put on hold 
because the attendants can no longer 
contact the reservation application. Once 
the application is restarted on the backup 
server, the attendants can restart checking 
in the passengers on standby. Still the 
reservations for the passengers checked 
in prior to the failure are not lost since 
they had been saved in the backup-server 
database.
It must be possible to combine 
high availability with disaster 
recovery
High availability requires the presence of a 
synchronous replication system between 
two servers. Both servers must be placed 
on the same LAN for two reasons. The 
first reason is that the LAN’s bandwidth 
and latency determines the synchronous 
replication performances. Secondly, having 
two servers on the same LAN ensures 
service IP address switchover when the 
application is restarted on the backup 
server.
Disaster recovery and high availability 
can be ensured simply by extending the 
same LAN with fiber optic cable in two 
geographically remote computer rooms. 
Thus, it is possible to simultaneously ensure 
application high availability and disaster 
recovery with two servers in two computer 
rooms and a real-time data replication 
solution across the network.
When data must be replicated via a low-
speed WAN data replication must be 
asynchronous, but the asynchronous 
replication system does not offer high 
availability: data is lost in case of failure. 
Asynchronous replication must be 
compared to a remote backup via the 
network and not to a high-availability 
system.
Does the high availability 
solution resists to "split brain" 
without corrupting data?
Split brain occurs in situation of network 
isolation between two servers. Each server 
becomes primary considering that the 
other has failed and runs the application. 
During split brain, some high-availability 
solutions can corrupt database with two 
active applications on the same database.  
It is necessary to avoid the double 
execution by testing an external network 
equipment acting as a witness between the 
two servers.
04 - High Availability (HA) Choosing a Solution
```

## Page 5

```
• High availability architectures and best practices 
• Software clustering vs hardware clustering
• Shared nothing vs shared disk cluster
• Virtual machine HA vs application HA
• Byte-level file replication vs block-level disk replication
• Synchronous replication vs asynchronous replication
• Software high availability cluster vs fault-tolerant system
• How to implement redundant servers with a simple software 
(Windows/Linux)?
• Heartbeat, failover and quorum in a Windows or Linux cluster
• How a virtual IP address works (Windows/Linux)?
• What is the RTO / RPO of a SafeKit high availability cluster?
• Alternative to Microsoft NLB with SafeKit network load balancingM
ore information on our web site
1. The ideal product for a software publisher
Harmonic, TV Broadcasting:“
SafeKit is the ideal application clustering solution for a software publisher looking for a simple and economical high availability software. 
We currently have more than 80 SafeKit clusters worldwide on Windows with our critical TV broadcasting application through terrestrial, 
satellite, cable and IP-TV. SafeKit implements the continuous and real-time replication of our database as well as the automatic failover of our
application for software and hardware failures. Without modifying our application, it was possible for us to customize the installation of 
SafeKit. Since then, the time of preparation and implementation has been significantly reduced.”
2. The product very easy to deploy for a reseller
NOEMIS, value added distributor of Milestone video surveillance solutions:
“SafeKit by Eviden is a professional solution making easy the redundancy of Milestone Management Server, Event Server, Log Server. The 
solution is easy to deploy, easy to maintain and can be added on existing installation. We have assisted integrators to deploy the solution on 
many projects such as city surveillance, datacenters, stadiums and other critical infrastructures. SafeKit is a great product, and Eviden 
provides great support. Happy to help if you have any questions.”
3. The product to gain time for a system integrator
Atos, BU Transport:
“SafeKit is a simple and powerful product for application high availability. We have integrated SafeKit in our critical projects like the 
supervision of Paris metro lines (CCR, centralized control rooms). Thanks to the simplicity of the product, we gained time for the integration 
and validation of the solution and we had also quick answers to our questions with a responsive Eviden team.”
3 partner testimonials
Thanks to this analysis using the example of the airport, it is now clear that a simple data replication 
solution is not a high-availability solution. We have seen the pitfalls in case of server failure and we advise 
the reader to check that a replication solution actually meets the high-availability criteria that we have just 
explained.
Many products make out that they perform data replication and offer high availability of applications, 
whereas in reality they only implement data replication and are very incomplete in terms of recovery in 
case of failure. With this type of product, an IT department insidiously believes it has a redundant high-
availability solution. But then it discovers the product’s limits the day a problem, such as a server blocking in 
its booting process, turns into a generalized crisis, with a day of service unavailability for an entire airport! 
SafeKit is the ideal solution for high availability of critical applications. It is a comprehensive and simple 
product. It was chosen after the crisis at the airport and has advantageously replaced the competitor's 
solution that had led to the critical situation described in this document.
High Availability (HA) Choosing a Solution - 05
```

## Page 6

```
1
0 reasons to choose the SafeKit clustering 
software
1. Software-only high availability solution
Eviden SafeKit is a software-only high availability solution. This solution secures easily and quickly the 24x7 operation of your critical 
applications. While traditional high availability solutions are focused on the hardware failover of physical servers, SafeKit has chosen to 
focus on the hardware and software failover of critical applications.
2. High availability which targets all types of failures
The unavailability of an application can be due to 3 types of problems:
• Hardware and environment: including the complete failure of a computer room (20%).
• Software: regression on software update, overloaded service, software bug (40%).
• Human errors: administration error and inability to properly restart a critical service (40%).
SafeKit addresses these issues, which are all essential to ensure the high availability of critical applications.
3. The 3 best use cases of software clustering
After over 20 years of 24x7 experience, SafeKit is the preferred clustering solution on the market in three cases:
• A software publishing company can add SafeKit to its application suite as a software OEM high availability option.
• A distributed enterprise can deploy a high availability solution on standard hardware without the need for or IT skills.
• A data center can provide high availability for multiple applications with a uniform solution on Windows or Linux and with load 
balancing, real time data replication and failover between two remote sites.
4. Unique on the market: 3 products in 1
Traditionally, three different products are necessary to create an application cluster:
• load balancing network boxes,
• disk bays replicated synchronously on a SAN for data availability, 
• high-availability toolkits for application failure recovery.
SafeKit offers these three features within the same software product.
To further reduce implementation costs, SafeKit runs on your existing physical or virtual servers and with the standard editions of OS and 
databases: Windows, Linux, Microsoft SQL Server, Oracle, Firebird, MariaDB, PostgreSQL or other databases or flat files… and even with 
Windows editions for PCs!
5. A solution suited for Cloud environments
Application high availability with SafeKit can be deployed in AWS, Azure and Google clouds as well as on premise on physical or virtual 
machines. Redundancy of Docker applications is also supported.
6. Full virtual machines replication and failover
SafeKit also offers replication and failover of full virtual machines between two active Hyper-V or KVM physical servers. The solution is 
simple and economical because it requires no shared disk.
7. Plug and play deployment of a software cluster
Once a failover module is configured and tested for an application, deployment requires no specific IT skills. Just install the application, the 
SafeKit software and the failover module on two standard Windows or Linux servers.
8.  Rich choice of application integration inside a software cluster
SafeKit proposes different types of software clusters. Cluster configuration for a given application is rich and is made with one or several 
application modules. SafeKit proposes mirror modules (primary/secondary with replication and failover), farm modules (network load 
balancing and failover), and mixed of several modules that can be implemented on the same cluster or on different clusters. 
A module is configured with the server IP addresses for heartbeats, the virtual IP address of the cluster, the load balancing rules for a farm 
module, the file directories to replicate for a mirror module, the hardware and software failure detectors and the service to restart in case of 
failure.
06 - High Availability (HA) Choosing a Solution
```

## Page 7

```
9. User-friendly administration to avoid human error
SafeKit provides a centralized administration web console. An administrator can remotely monitor status of applications on different 
clusters and act with simple buttons (start, stop). 
You can test SafeKit for free. In less than 1 hour, you can implement your first software cluster on two virtual or physical machines thanks to 
the administration console.
10. Synchronous replication for transactional applications
SafeKit’s synchronous real time replication function strengthens high availability and prevents data loss. With this mechanism, a data 
committed on a disk by a transactional application is replicated on the secondary machine.
Application servers can be located in geographically remote computer rooms through an extended LAN to withstand the loss of a full 
room.
High Availability (HA) Choosing a Solution - 07
```

## Page 8

```
A
bout Eviden
Eviden IAM is the Identity and Access Management (IAM) 
software suite of Eviden.E
viden IAM is the European leader in identity and access ma-
nagement with a presence which is growing rapidly beyond 
Europe, particularly in Japan and the US.M
ore than 5,000,000 users in more than 900 organizations 
throughout the world connect to their companies every day 
and manage their access rights with Eviden identity and ac-
cess management solutions.C
T-210325-JR-HIGH-AVAILABILITY-CHOOSING-SOLUTION-EVIDEN-ENG
© Eviden. All products, brand names, service marks, trademarks and other names 
mentioned in this document are proprietary to their respective owners and are 
protected by applicable trademark and copyright laws. Eviden reserves the right 
to modify the characteristics of its products without prior notice.F
or more information: SafeKit web site
```