TFTP Balancing with NetScaler

TFTP Balancing with NetScaler -

UPDATE (April 2014) load - My colleague has just published an article with the new optional NetScaler 10.1 TFTP LB to check it out

I published an article entitled! "Load Balancing TFTP - Trivial Anything But" a little over a year ago. And since then, it has received over 10,000 visits and it was one of the most popular articles I've ever written in terms of comments and emails offline I receive about the article. If you have not had the chance yet to read that article (or comments), please start there. Because it is essentially a follow-up article.

Now, if you read that article (or the associated white paper is much better written) you will see that I have provided a few options to balance the load of the TFTP service NetScaler with (especially options 5, 6 and 7). What I would do in this article is drawn on these very generalized options I provided in the last article. Because, to be honest with you, this article does not really do justice to the NetScaler side of the house. And with a growing number of NetScaler appliances deployed there, this is a perfect time to return to this subject.

Before I jump into the specific options to balance the TFTP load with NetScaler, I have to give credit for 99% of the content of this article to my friend and colleague, Mike Skowronski . As for me, Mike is an architect of the team of Consulting Americas. But Mike specializes almost exclusively in our practice of the network - and is one of the best NetScaler guys we have in society, in my opinion. again thank you Mike! Let's get to it ...

Options for balancing the load with TFTP NetScaler

1. USIP

• what is this? USIP the NetScaler mode keeps the original source IP address and it goes back-end server.
• What are the advantages?
  • Easy to set up.
  • Works with L2 and L3 adjacency.
  • Can use all algorithms LB available for the protocol.
• What are the disadvantages?
  • In-1-arm topology, TFTP servers default gateway must point to the NetScaler SNIP or MIP.

2. L2 DSR

• What is this? L2 DSR we add the VIP TFTP NetScaler to the loopback interface on all final TFTP servers. The TFTP server then responds directly to the client, bypassing the NetScaler on the return.
• What are the advantages?
  • Quick.
• What are the disadvantages?
  • As its name indicates, works only with L2 adjacency to the TFTP server.
  • slightly complicated configuration on the NetScaler.
  • Requires changes to the loopback interface on top on each TFTP servers.
  • limited set of load balancing algorithms.

3. L3 DSR

• What is this? DSR mode L3 we need to enable IP tunnel on the final TFTP servers. Otherwise, as L2 DSR.
• What are the advantages?
  • Works with L3 adjacency.
• What are the disadvantages?
  • complicated configuration on the NetScaler.
  • Requires changes to the loopback interface on top on all TFTP servers.
  • Requires changes to the network stack on the server TFTP enable IP tunneling.
  • limited set of load balancing algorithms.

4. L2 DSR session without

• What is this? In this mode, we disable session tracking on the TFTP VIP. Otherwise, as L2 DSR.
• What are the advantages?
  • Very light and extremely fast.
• What are the disadvantages?
  • slightly complicated configuration on the NetScaler.
  • Requires changes to the loopback interface on top on all TFTP servers.
  • works only with L2 adjacency.
  • very limited set of load balancing algorithms.
  • works only in topology 2 arms.
  • Each TFTP service can be linked once to a single VIP TFTP.

5. L3 DSR session without

• What is this? In this mode, we disable session tracking on the TFTP VIP.
• What are the advantages?
  • Lightweight and fast.
  • No changes required on the back-end servers.
• What are the disadvantages?
  • very limited set of load balancing algorithms.
  • works only in topology 2 arms.
  • Each TFTP service can be linked once to a single VIP TFTP.

6. DS DSR

• What is this? Only available in NetScaler v10, we can use the Differentiated Services field in the IP header to encode a TFTP VIP ID to support DSR.
• What are the advantages?
  • Works with L3 adjacency.
• What are the disadvantages?
  • slightly complicated configuration on the NetScaler.
  • complicated configuration on the TFTP backend servers. Back end TFTP servers must be configured to map the DS ID VIP TFTP correct when responding to the client.
  • limited set of load balancing algorithms.
  • DS field is 6 bits so limited to 63 vservers (0 is not a valid ID) that can be followed by a back-end server.

7. Same-Site GSLB

• What is this? In this mode, we set up multiple sites on the same GSLB NetScaler and each site has one VIP TFTP linked to a single TFTP server. We then use the GSLB load balancer to balance the load between sites and therefore accordingly to get the load balancing TFTP service.
• What are the advantages?
  • Works with L3 adjacency.
• What are the disadvantages?
  • Requires TFTP clients are configured to use FQDN instead of IP addresses.
  • requires a DNS infrastructure is in place.
  • Complicated NetScaler configuration.
  • This could be considered a kind of piracy by some purists

So there you have it -. a much more complete list of options if you NetScaler available. Now I'm sure you are wondering ... if the network configuration and complexity do not matter, what option would you recommend? I asked Mike the same question and he said he is a fan of one of DSR session without methods.

We hope this information helps. If you have comments or questions, we please to drop us a line below. I will do my best to answer every comment / question and Mike will also in response to comments / questions below as well.

Cheers, Nick

Nick Rintalan, Senior Architect, Enterprise Architecture, Americas Consulting

Tweet