Bookmark and Share

2011.07.04 - Migration to IPv6: There’s no Hurry!

Although the lack of IPv4 addresses is threatening the Asia Pacific region, IPv6 Day on June 8, 2011, helped build excitement for second generation IP addressing. It also created a sense of urgency to migrate your entire information system to IPv6. In reality, migrating to IPv6 is neither urgent, nor required.
 
1. IPv4: the reality of its limits
It is a known fact that IPv4 is reaching saturation. APNIC, which manages addresses in the Asia Pacific region, announced in April 2011 that the IPv4 address pool was fully depleted. Excluding Africa, all other continents have almost exhausted their pools. But what addresses are we talking about? Public addresses, in other words those that are used to identify a system from the outside. Businesses’ internal networks, which each have their own internal addressing plan, are not directly affected by the shortage.
 
2. One (Front-End) public IPv6 address is enough
This is the logical conclusion of the previous point: businesses must be visible in IPv6 by outside systems, or they will lose traffic and possibly even potential customers. In practice, businesses simply need to have a public IPv6 address and set up a gateway, such as a load balancer, that can translate IPv6 addresses to IPv4, so that they can continue to use IPv4 on their application servers.
 
3. Even so, are new applications or infrastructure necessary?
No, at least not urgently and inevitably. That is the advantage of address translation. Adding a gateway allows you to keep your infrastructure as is, even if its components do not support IPv6 addressing. The same is true for applications, which continue to run internally on IPv4. So there is no need to revise your investment plan: switching and migrating information system components one-by-one (to solutions that will manage IPv6 natively) will allow you to migrate everything to IPv6 naturally and gradually.
 
4. Is address translation intrusive in the information system?
No. With a layer 7 load balancer, which operates on already established connections between the user and the Web application, translation is fully transparent. Once IPv6 traffic goes through the load balancer, it arrives at the application servers in IPv4.
 
5. Address translation: a dedicated or pooled tool?
It all depends on the legacy infrastructure. For businesses that already have load balancers that can manage IPv6 to IPv4 address translation, 3 configuration lines will suffice. For businesses with load balancers that are not compatible with the new protocol and/or that only operate at layer 4, an entry level load balancer – dedicated to IPv6 to IPv4 addresses translation – can be placed at the front of the network.
Note: managing coexisting IPv6/IPv4 addresses is by no means the primary function of a load balancer. Businesses can take advantage of the load balancer’s other functions: load balancing and high availability, performance optimization, acceleration and application security.
 
6. Does IPv6 to IPv4 address translation slow down transactions?
With a layer 7 load balancer, connections with the customer and the server are independent, so translation does not slow down transactions. The only possible reason for a slow down is the IPv6 protocol itself, the management of which is more resource-intensive for IT equipment. With modern hardware processing power, however, the additional processing time is only several microseconds.
 
7. When should a business have a fully IPv6 information system?
With bilateral address translation, a fully IPv6 information system is not required. Businesses can gradually replace IPv4 only compatible software and hardware with newer models that support IPv6. Some businesses may make the transition in a few years, while others may still be using IPv4 fifty years from now!

© 2016 Haproxy Technologies, LLC. All rights Reserved. E-mail: info@haproxy.com