What is a CoreLab Cluster?
This is the first high-availability solution equipped with an event-based (Interrupt) failure detection
technique developed by CoreLab.
This is a solution that can be applied to meet the customer’s IT system requirements as a single solution
for high availability (HA), disaster recovery (DR), and single server recovery (SSP).
What are the product features of the CoreLab Cluster?
Compared to the existing polling method (structure that receives resource responses at regular time intervals),
the failure detection time is more than 50% faster.
Since it receives responses almost in real time at the speed of the OS response, it can reduce downtime.
Windows MSSQL HA environment – Average failover time 10 seconds
Linux MariaDB HA environment – Average failover time 5 seconds
Since it can be operated on a single screen using the intuitive UI web console,
it is very easy to use compared to other solutions.
When maintaining the solution, there is no downtime because it updates by replacing only the modules
that require patches, rather than shutting down, deleting, rebooting, and reinstalling the operating service.
What is GS (Good Software) certification?
As of April 3, 2025, it has GS (Good Software) certification level 1.
The product supports x86 (Windows, Linux).
What exactly is an event-based (Interrupt) fault detection technique?
There are two ways to detect failures in APPs running on the OS: the traditional polling method
and the event method developed by us.
Polling method
It sends commands to the target resource at regular intervals and receives responses.
It usually operates in a cycle of 5 to 10 seconds.
The number of times is increased to increase the accuracy of the response. It is used about 3 times by default,
and if a response comes within that time, it is considered normal.
Usually, there is a random waiting time of 10 to 30 seconds, and in some cases, it takes close to 1 minute.
Event (Interrupt) method
It has its own event handler and recognizes resource status changes in the OS in real time.
Compared to the polling method, The waiting time is reduced, supporting fast fault detection
(average detection speed is 5 to 10 seconds).
What are the main features of the product?
– Supports resource failure processing using event-based (Interrupt) detection techniques rather than
periodic monitoring (Polling) of the target resource.
– Container-based resource implementation allows users to control/operate on a single screen.
– Displays resource dependencies in levels.
– Provides block-based point-in-time recovery function.
– Configures level-based dependencies in the Drag and Drop method.
-Supports multi-nodes such as N:1, 1:N, Active-Active (mirror and shared environment).
– System pre-hang-up response possible through performance monitoring (CPU, Memory).
– Supports Virtual Mac, Virtual NetBios Name, Replace IP.
– Provides user interface based on web GUI.
– Uses one single port for communication between redundant nodes.
– Provides AnyTime scheduler function for repetitive tasks (built-in S/W function).
– Provides alarm function by setting SMTP and Telegram in case of failure.
– Provides separate program for integrated control when there are multiple sets of redundant servers.
– Data is replicated at the block level for data replication efficiency.
– Network fencing and SCSI fencing are supported when Split-brain occurs to ensure data
consistency.
What are the main configurations that the product supports?
[High Availability: HA]
1:1 local replication, 1:1 shared configuration
N:1 configuration (Active – Active -Standby)
– Configure multiple tasks as A-A and group 1 server as a standby server
Here, A-A is a different concept from Oracle RAC (RAC configures the same task as A-A)
[Disaster Recovery: DR]
1:1 DR configuration (remote disaster recovery configuration for disk volumes)
[High Availability + Disaster Recovery: HA+DR]
1:1:1 HA + DR configuration
Mirror HA + DR: MDR)
Storage HA + DR: SDR)
[Single Server Recovery: Single Server Protection]
Supports monitoring and recovery of APP in StandAlone mode on a single server
What exactly is 1:1:1 HA+DR?
It literally combines high availability and disaster recovery. In preparation for daily failures
(system failure, network failure, service failure] and in the event that the main center is paralyzed due to
natural disasters such as fire, power outage, war, or earthquake, each HA and DR function
has been implemented so that they can be used as an integrated solution rather than as separate solutions.
It is usually expressed as MDR (Mirror + DR), and there is also SDR using storage.
Local HA functions as a high-availability cluster between each other during normal times,
and when the function does not work, it operates the service on the DR server, which is a backup center,
through a separate DR declaration.
Since this is a disaster situation, the goal is to resume the service within 4 hours at the most after
separate preparatory work (network and DNS, etc.) at the DR center.
HA provides business continuity by resuming the service within 1 minute between A-S
unless there are special circumstances, and DR is a bit more complicated and supports
resuming the service within 4 hours.
However, data consistency is guaranteed in real time during normal times, and the replication network
between HA servers is at least 1Gbps or more and the DR server has a minimum bandwidth of 100Mbps
or higher is recommended.
How do I maintain my product?
Due to the nature of the software, binary replacement is supported in case of bugs or patches.
Therefore, product patches can be made without interruption.
There is no downtime due to patches for server redundancy solutions.
How many licenses should I purchase based on the configuration method?
1 License – Single Server Protection (1 Node)
2 License – HA or DR (2 Node)
3 License – HA+DR (3 Node)
Please contact us for further details.
How to purchase a product?
[Sales]
+82-70-8847-2527, Sales@corelab.cc
[Technical]
+82-70-8847-2527, Support@corelab.cc