13.Netscaler Load balancing persistency and method

Summary

The video script explains load balancing algorithms, focusing on two common methods: Round Robin, which distributes requests evenly across servers, and Least Connection, which sends requests to the least busy server. It also introduces the concept of persistency to ensure a user's session remains on the same server for a consistent experience. Methods like Source IP and HTTP cookies are discussed for maintaining session continuity. The script concludes with a demonstration of configuring these settings on a NetScaler to observe load distribution and persistency in action.

Takeaways

• 🔄 **Round Robin Algorithm**: Distributes client requests sequentially across servers without checking server load.
• 🔗 **Least Connection Method**: Directs incoming requests to the server with the fewest current connections, ensuring even load distribution.
• 📈 **Load Balancing Methods**: Includes round robin, least connection, least response time, least bandwidth, hashing, and custom methods.
• 🔒 **Persistency**: Ensures a user's session remains on the same server throughout a transaction, preventing session loss or confusion.
• 📍 **Source IP Persistency**: Keeps requests from the same IP address directed to the same server for a specified time interval.
• 🍪 **HTTP Cookie Persistency**: Uses browser cookies to maintain a session on one server, which can be lost if the browser is closed.
• 🛠️ **Configuration of Load Balancer**: Involves setting the load balancing method and persistency to meet application requirements.
• 🔄 **Backup Method**: A secondary method to use if the primary load balancing method fails.
• 📊 **Monitoring Load Distribution**: Observing the behavior of the load balancer to ensure it is distributing load as intended.
• 🔧 **Adjusting Configuration**: Saving and testing changes to the load balancer configuration to observe the impact on load distribution.
• 🛡️ **Importance of Persistency**: For applications that require session continuity, persistency is crucial to prevent issues with transaction integrity.

Q & A

What is a load balancing algorithm?

A load balancing algorithm is a method used to distribute network or application traffic across multiple servers to ensure no single server bears too much load, thus improving responsiveness and availability.

Can you explain the Round Robin load balancing method?

Round Robin is a load balancing method where requests are distributed sequentially across servers. The first request goes to the first server, the second to the second server, and so on, cycling back to the first server once the list is exhausted.

What is the Least Connection method in load balancing?

The Least Connection method checks the number of existing connections on each server and sends incoming requests to the server with the fewest connections, aiming to evenly distribute the load.

What is the purpose of persistency in load balancing?

Persistency ensures that a user's session or transaction is consistently handled by the same server, preventing issues that may arise from a session being handled by multiple servers without awareness of the previous interactions.

How does Source IP persistency work in load balancing?

Source IP persistency directs all requests from a specific IP address to the same server for a defined period. This ensures that a user's session is maintained on one server, preventing session-related issues.

What is HTTP cookie persistency and how does it differ from Source IP persistency?

HTTP cookie persistency uses a cookie sent to the user's browser to ensure that all requests from that browser session are directed to the same server. It differs from Source IP persistency in that it is browser-based, not IP-based, and the session is lost when the browser is closed.

Why is it important to configure a backup method in load balancing?

A backup method is important in case the primary load balancing method fails. It ensures that the load balancing process can continue seamlessly without significant disruption to the service.

What are some other methods for load balancing besides Round Robin and Least Connection?

Other load balancing methods include Least Response Time, which sends requests to the server with the shortest response time, List Bandwidth, Hashing Method, and Custom Method, which can be tailored to specific needs.

How can you configure load balancing and persistency settings on a NetScaler?

On a NetScaler, you can configure load balancing and persistency by going to the 'Method' option under the 'Whip' settings, where you can select the load balancing method and set persistency rules, such as Source IP or HTTP cookie.

What is the significance of setting an interval for persistency?

Setting an interval for persistency is crucial to prevent overloading a single server with all connections. It ensures that the load is distributed over time, maintaining the balance and preventing server overload.

Outlines

00:00

🔄 Load Balancing Algorithms and Persistency

This paragraph introduces the concept of load balancing algorithms, specifically focusing on two common methods: round robin and least connection method. Round robin distributes client requests sequentially across servers, while the least connection method sends new requests to the server with the fewest existing connections. The paragraph also touches on other methods like least response time, list bandwidth, hashing, and custom methods. A key aspect of load balancing discussed is persistency, which ensures that a user's session remains on the same server throughout a transaction, preventing issues like an empty cart during the payment process in an e-commerce scenario. Persistency methods like Source IP and HTTP cookies are mentioned, which help maintain the session continuity on a specific server for a defined period.

05:00

🔒 Configuring Persistency in Load Balancing

The second paragraph delves into the practical application of persistency in load balancing. It explains the importance of setting persistency to ensure that a user's session consistently connects to the same server for a certain duration, which is crucial for applications that require continuity in a session. The speaker demonstrates how to configure persistency in a load balancer, such as a Netscaler, by setting a time interval during which a user's session should remain on the same server. The paragraph also highlights the fallback methods that can be used if the primary load balancing method fails, emphasizing the flexibility and reliability of load balancing configurations.

Keywords

💡Load Balancing

Load balancing is the process of distributing network or application traffic across multiple servers to ensure no single server bears too much demand, thus avoiding overload. In the video, it is the central theme, with various algorithms discussed to achieve this, such as round-robin and least connection methods, to enhance efficiency and reliability of services.

💡Round Robin

Round Robin is a load balancing algorithm that cycles through a list of servers and distributes client requests in turn to each server. It is mentioned in the script as a method that does not check server load but simply rotates requests, ensuring a fair distribution of traffic without considering the current load on each server.

💡Least Connection Method

The least connection method is an algorithm that sends incoming traffic to the server with the fewest current connections. It is highlighted in the script as a more intelligent load balancing approach compared to round robin, as it takes into account the existing load on each server to make a more balanced distribution decision.

💡Persistence

Persistence in load balancing refers to maintaining a user's session on the same server throughout their interaction to ensure continuity and consistency of experience. The script explains its importance, especially in scenarios like e-commerce where a user's shopping cart should be consistent across the session, and mentions techniques like Source IP and HTTP cookies to achieve persistence.

💡Source IP Persistence

Source IP Persistence is a method of ensuring that all requests from a particular IP address are directed to the same server for a specified duration. The script uses this as an example to illustrate how persistence can be configured, ensuring that a user's session remains consistent even when load balancing is in effect.

💡HTTP Cookie

HTTP Cookie persistence is a technique where a cookie is set in the user's browser to direct all requests from that session to a specific server. The script explains that this method is browser-based and will ensure that a user's session remains on the same server until the browser is closed or the cookie expires.

💡NetScaler

NetScaler is a product mentioned in the script that is likely a load balancer or similar device used to manage traffic distribution. It is used as an example to demonstrate how to configure load balancing methods and persistence settings to achieve desired traffic management outcomes.

💡Load Balancing Methods

The script discusses various methods for load balancing, including round robin, least connection, least response time, least bandwidth, hashing, and custom methods. These methods are the different algorithms or rules that can be applied to distribute traffic across servers in a load balancing setup.

💡Session

A session in the context of the script refers to a series of interactions between a user and a server, such as adding items to a shopping cart on an e-commerce site. The script emphasizes the importance of maintaining session consistency through load balancing to ensure a seamless user experience.

💡E-Commerce Portal

The script uses the example of an e-commerce portal to illustrate the practical application of load balancing and persistence. It explains how a user's shopping and payment process should ideally be handled by the same server to maintain session data and provide a smooth shopping experience.

Highlights

Load balancing algorithms distribute network traffic based on certain conditions.

Round-robin method distributes requests sequentially across servers without checking load.

Least connection method directs traffic to the least loaded server based on current connections.

Other load balancing methods include least response time, list bandwidth, hashing, and custom methods.

Persistency ensures a user session remains on the same server for the duration of a transaction.

Source IP persistence directs traffic from the same IP address to the same server for a defined time.

HTTP cookie persistence uses browser cookies to maintain a session on one server.

Load balancers can be configured to switch methods if the primary method fails.

Demonstration of load balancing in action, showing how traffic is distributed.

Setting up persistency in a load balancer to ensure consistent server allocation for a session.

The importance of defining a time interval for persistency to prevent overloading a single server.

Load balancing configurations can be changed to suit different application requirements.

Backup methods can be set in case the primary load balancing method fails.

NetScaler is used as an example of a load balancer where methods and persistency can be configured.

A live demonstration of how persistency works in a load balancer, showing session consistency.

Load balancing is crucial for maintaining application performance and server health.

Transcripts

load balancing uh algorithm so what is

load balancing algorithm so load

balancing algorithm is nothing but based

on the conditions you are doing the load

balancing for example let's say round

robin round robin is uh first if if you

have 50 request from the client the

first request it may go to the first one

second it may go to the second third

minute go to the first one so randomly

it will throw so it won't check anything

randomly it will distribute across your

back end okay so that is called round

robin and the second one is list

connection method least connection

method means the net scaler will check

whether the back end is having how many

connection for example This Server is

having 50 connection and the second

server is having only 30 existing

connection then the next incoming

packets right it will send you to the

second server unless it matches you know

unless it identifies uh the load are

distributed equally in the back end so

these are the two most common method we

use to do the load balancing one is

round robin which is automatically it

will distribute another one is list so

it will check in the back end which is

at least loaded and accordingly it will

distribute there are other method least

response time list bandwidth hashing

method custom method so this is you can

read about it so in a nut cell this is

how uh this is the load balancing

algorithm it will use to distribute so

you may see or where I have to configure

this one so I'll tell you something

before going to that part there is one

more thing called persistency

so this is very important so when you do

load balancing right there is something

called persistency persistence is

nothing but for example let's say you're

initiating a connection for example

you're trying to or you know access one

e-commerce portal maybe you're doing

some shopping okay so while you're

adding all the items to the shopping

cart then the request may go to the

server one and server will process

everything related to your items okay

when you click on payment so let's say

the connection may go to the another

server so the second Server doesn't know

any idea what was happened in the

earlier session right because forget

about

the advanced like in the back end you

may have a DB that may have the

information Etc forget about that

consider a very simple processor where

you don't have anything back in so in

this scenario server one processed

everything related to your purchase but

when you are paying the bill the

connection may go to the server too

because you are doing the load balancing

so server 2 doesn't have any idea it may

show empty cart or it may so value as

zero so in that scenario what you have

to do is you have to do something called

persistency so that your session for

example your user one so user one should

be connected to server run unless until

he completes his complete transaction

you got my point so everything should

sort from the server one itself unless

and until you know he initiate a new

connection so in his existing connection

it should go everything to one

particular server for that we need to

configure something called persistency

okay

so persistency there are n number of

method for the persistency the most

famous one is Source IP Source IP is

nothing but for example let's say the

user is coming from 192 168 100.5 so I'm

telling that uh to the net Skiller any

request coming from this particular IP

address for the next five minutes or

next 10 minutes my session should go to

server one itself so let's say you have

another user which is coming from

another source IP let's say

10.168.5.5 so for that the net skill

will always send the communication to

five minutes sorry server two but there

is an interval okay only for five

minutes or only for 10 minutes because

if you do not set interval what will

happen eventually all the connection May

load on the first server or on the

second server it may choke and it will

have issues so you need to Define till

how many how much time you want to send

the communication to the one server

itself so this is very important when it

comes to persistency HTT Source IP is

the one method which you are setting for

the persistency another one is HTTP

cookie http cookie is nothing but it is

browser based for example let's say you

are trying to open a browser and you are

accessing your load balanced site okay

so netscaler will send a cookie to that

particular browser session and you will

say all the browser session coming from

this particular browser session it

should go to only one server okay so if

you close the browser and if you

relaunch the browser again then it may

send it to another server why because

your your cookie will get lost because

when you close the browser so cookie

insert is the browser based and Source

IP is based on the source IP so let's

see in demo you will get to know

so see now it is Distributing right see

now it is Distributing let's set the

cookie uh this persistency method and

we'll see the behavior okay I'm going to

the net scalar so in your netscalers uh

we go to the Whip and here you have one

option called method right so under

method you have to option to change the

method round drop in or List connection

Etc here you will see the method to

change the load balancing so by default

it will be uh list connection you can

change it to round robin etc etc so you

have a backup method also so let's say

due to some XYZ reason this method is

fails what is the next method you should

take so this is related to method on

which you do the load balancing now

coming to the persistency so this is

very important so there is as of now

there is no persistency that is why it

is Distributing load across all the

servers so let's set a persistency and

say for the next two minutes my server

my session should go to the same server

okay so I'll click OK and I will click

done

so now let's save the configuration

if everything is good then let's see the

behavior I'm going to copy this URL and

I'm going to open a new session

see it is connected to server one let's

refresh see for the next two minutes it

won't change it will go to the server

one itself

this is the beauty of persistency

because some application uh you know

while you're doing the load balancing

which will have its own requirement that

or or you know some server some session

should served in the same server okay

you should not go to the other server so

you can use this method and you can

overcome that so this is related to load

balancing of HTTP and uh