Joaquin Attanasio

Joaquin Attanasio

Business Intelligence Consultant | Microstrategy Expert | Data Specialist

Other Articles:

MicroStrategy Fences

Share on linkedin
LinkedIn
Share on facebook
Facebook
Share on twitter
Twitter
Share on whatsapp
WhatsApp

Hello everyone, how are you? After several weeks of absence, here I am again in our #BestInMicro space.

With the latest update of the platform (11.3.7), the management of Fences in Workstation has been enabled. And while many were jumping for joy, most people asked me ‘What do you eat that with’? So, now we have a topic of conversation. Today I’m going to explain what MicroStrategy Fences are.

As always, first things first. It is important to understand what a cluster is, what it is for and what types we have.

Clusters

Basically, a cluster is a set of servers working together, either to:

  • Improve performance (by adding physical processing and storage capacity)
  • Ensure high availability (i.e., if a server goes down for whatever reason, the service is not cut off)
  • Provide facilities when extending or expanding our environment (always talking about capacity, not implementation).
  • Manage workload, assigning specific nodes to each project.

Today we are going to focus on this last point. It is common to find cases where a project has a much higher demand or volume of data than the rest, or where certain people need processing priority.

We can see the servers that are associated in cluster in the properties of the Intelligence server, or also in the menu administration > cluster nodes

Types of clusters

We can define clusters for both the intelligence server and the web server. In the following image, for example, we can see how the architecture is in both cases, obtaining the benefits that we mentioned in the previous step.

However, there are certain scenarios where high availability and thread distribution are not a benefit in a cluster. This is given when we talk about an Asymmetric cluster.

Basic or symmetric clusters are practically copies of different servers working in parallel. An asymmetric cluster is two or more servers connected, but with different distribution of services, or in our case, projects (as you will see in my work of art below).

So, what is the advantage of having an asymmetric cluster? That we can distribute the projects according to their load, availability, accessibility, etc… In this type of cluster, the focus is on load distribution.

And this is where the functionality we have been talking about comes in, in force since version 10.4, but now brought to the Workstation: Fences.

This is a feature that, logically, exists in the then clustered systems, and is used to reserve nodes for the exclusive use of users or for the execution of different tasks.

Types of Fences

In other words:

User Fences: used to process requests from a list of specified users or user groups.

Task Fences: used to run scheduled jobs with a subscription (except for ‘run now’, in case there are any creative types out there looking for the trap!)

The following diagram shows a clear example of how a distribution of nodes can be made, using Fences, with some examples of how the workload is distributed.

Limitations

Everything very nice and pretty with this load distribution capability, because it really brings certain advantages, especially in terms of performance, when facing users is one of the things that most often make a difference (You know, who does not know any user who complains that the tool is slow…), but there are important limitations to consider. I recommend you to read this article where it details some limitations that it is good to consider when using this functionality, but next week I will bring another article where I will get into detail to know if it is really convenient or not to use the functionality, and logically to understand its impact.

So that is it for today! Let me know if you knew or not this functionality, if you used it, and even if you would like us to test it in a video.

I hope you found it useful and see you next time!

References

Deja un comentario

Your email address will not be published. Required fields are marked *

Business Data Master Logo

No te pierdas el

WEBINAR
Gratuito

Explicaremos en detalle los contenidos y objetivos del Business Data Master

29/11/2021

18:30 (GTM+1)

Online

BUSINESS DATA MASTER

* Tu información será utilizada exclusivamente para contactarte en relación al Business Data Master. No hacemos spam ni compartimos datos con terceros.