Forum Discussion

gorge_300427's avatar
gorge_300427
Icon for Nimbostratus rankNimbostratus
Apr 14, 2017
Solved

Why do we need a F5 LTM

Hello guys I am new to F5 . And I am looking to learn LTM . But I cant find any books or articles which explain why do we need a F5 in the first place. I know it can load balance efficiently , but that even a Router can. I do know that F5 monitors traffic of the servers which the router cant and it inspect , intercept traffic upto the appplication layer. I am looking to understand why did we do the transition from using a router to using a load balancer , and what all benefits does it give us apart from the ones I know. Scenarios would be of great help!

 

Thanks

 

application delivery
LTM
  • Kevin_Davies_40's avatar
    Kevin_Davies_40
    Apr 14, 2017

    The answer to your question is availability. The primary goal of the Local Traffic Manager is application availability far beyond what a router can possibly deliver. The F5 is not a router, far from it. It is a full proxy networking device with a default deny security posture. It can proxy in a few different ways but by far its layer seven capabilities and world class SSL performance are some of its features.

     

    Application Delivery as it is now known, is about making decisions in real time about incoming connections from a client. These decisions not straight forward. At a router level you are delivering packets much like a mail sorting system in a post office. If they are not home, you just leave the mail there. At an application delivery level you are reading the mail, deciding if the same sender should always go back to the same house, even to the same person in that house. Then if they are not home, is their someone else in the house who can accept the mail, if their is not is their another house available, if not, do they have a house in another state/country (GTM) and are they home there? At its core the F5 is about making sure their is actually a person (application) there to receive the mail and process it.

     

    When it comes to application delivery the house is our pool of services. Those services all serve up the same application, usually from different servers. So how do we deliver new connections to them? Are we going to loop through them or in application delivery world, choose them based on loading at layer four, layer seven and even processor utilization. Are we going to just check their availability with ICMP? application port with TCP? or actually send valid requests to the application itself at layer seven and determine availability on the response we receive. A common example is connecting to and SMTP mail server, having a conversation and then disconnecting. Another is logging into an FTP server, downloading a file and then disconnecting.

     

    This is what application delivery is all about. Ensuring your traffic reaches an application and that is always available. That you can scale your single server application to many servers and the BIG-IP can handle the unique challenges this can present. To ensure that in failure scenarios their is no downtime with your servers and even the BIG-IP's themselves having extremely reliable failover from a single pair to even eight of them working together in high availability. F5 Networks are world leaders in this area, their products can be found in nearly all sectors where application availability is a must.

     

    To have a better understanding have a read over the nuts and bots of load balancing. Then head over and try out some free training provided by F5 Networks!

     

3 Replies

Sort By
  • Hectorm's avatar
    Hectorm
    Icon for Nimbostratus rankNimbostratus
    Apr 14, 2017

    Well a router is a layer 3 device; While the BIGIP is a layer 7 device. This means that you can balance device base on content and layer 7. Also Once you learn how to use Irule the question you will be asking is why would anybody use a router to try to load balance. LTM is a real man in the middle device. this means that users do not connect to the servers. The user connects to the BIGIP and the connection terminates at the BIGIP. Then the BIGIP creates a new connection between the Bigip and the server. This means you can manipulate the data before it goes to the servers as well as when it comes back from the server. Now to learn about the BIGIP go to F5 university in the web, the training and videos are free and you can learn all the F5 can do for you.

     

  • Kevin_Davies_40's avatar
    Kevin_Davies_40
    Icon for Nacreous rankNacreous
    Apr 14, 2017

    The answer to your question is availability. The primary goal of the Local Traffic Manager is application availability far beyond what a router can possibly deliver. The F5 is not a router, far from it. It is a full proxy networking device with a default deny security posture. It can proxy in a few different ways but by far its layer seven capabilities and world class SSL performance are some of its features.

     

    Application Delivery as it is now known, is about making decisions in real time about incoming connections from a client. These decisions not straight forward. At a router level you are delivering packets much like a mail sorting system in a post office. If they are not home, you just leave the mail there. At an application delivery level you are reading the mail, deciding if the same sender should always go back to the same house, even to the same person in that house. Then if they are not home, is their someone else in the house who can accept the mail, if their is not is their another house available, if not, do they have a house in another state/country (GTM) and are they home there? At its core the F5 is about making sure their is actually a person (application) there to receive the mail and process it.

     

    When it comes to application delivery the house is our pool of services. Those services all serve up the same application, usually from different servers. So how do we deliver new connections to them? Are we going to loop through them or in application delivery world, choose them based on loading at layer four, layer seven and even processor utilization. Are we going to just check their availability with ICMP? application port with TCP? or actually send valid requests to the application itself at layer seven and determine availability on the response we receive. A common example is connecting to and SMTP mail server, having a conversation and then disconnecting. Another is logging into an FTP server, downloading a file and then disconnecting.

     

    This is what application delivery is all about. Ensuring your traffic reaches an application and that is always available. That you can scale your single server application to many servers and the BIG-IP can handle the unique challenges this can present. To ensure that in failure scenarios their is no downtime with your servers and even the BIG-IP's themselves having extremely reliable failover from a single pair to even eight of them working together in high availability. F5 Networks are world leaders in this area, their products can be found in nearly all sectors where application availability is a must.

     

    To have a better understanding have a read over the nuts and bots of load balancing. Then head over and try out some free training provided by F5 Networks!