Spotlight

Tony Hill shared his take on this practical problem based on his wealth of experience. “We had the same experience in Australia. It took 12 months before government agencies take IPv6 seriously. And then, they discovered they had so many apps to deal with. How they approached it was to divide the world between external and internal. The external world will be the first phase to operate in dual stack environment. For the internal world, it gets more complicated. It might require an internal audit.”

Hill also noted that network management tools can automate the audit process and quickly take stock of network components that lack IPv6 support with minimal time and manpower. This is particularly useful when managing large networks, as time and manpower costs saved can be significant compared to a manual audit.

“Although at layer 7, apps shouldn’t be impacted by layer 4 and 5, often times, coding can be suspect. Companies should have a test bed network where users can test their apps. This could be done on an internal IPv6 day.”

Yew Hock Meng understands the pain points well. He advocates early planning and migration. “It took us three years to plan, three months to test and three months to migrate. The important thing is to start planning and take stock of your network. Which one lacks IPv6 functionality? Taking the plunge into IPv6 takes months, but once you have a methodology, it is not so scary”.

Willy Sutrisno stressed the importance of ensuring the peaceful coexistence of both IPv4 and IPv6 networks. “Think of IPv6 coexisting with IPv4. IPv4 won’t just expire. For example, we have a company portal which does not need to be IPv6 capable as it is only for internal use. But if your system needs to be accessed from the Internet, then you should consider a transition as soon as possible.”

With multiple methods that companies can adopt to transit to IPv6 depending on their specific circumstances, Miwa Fujii advocates learning as much as possible before choosing a method. “There are so many different ways to transit: dual stack, NAT444, DS-Lite, and NAT64. Different ways are more applicable for different customers. So, learn as much as possible.”

Getting IT staff properly trained is part of the process to ensure a smooth transition to IPv6. Victor Tang thinks that companies should do this by taking advantage of the heavily subsidised training programmes available in Singapore. “IDA has taken the initiative to build IPv6 competency in Singapore. We provide IPv6 Forum certification courses. This allows a company’s engineers to appreciate the fundamentals, learn the different aspects of IPv6, and understand how to implement it in a network. We provide these courses with IDA funding of up to 70 percent. Companies can send their engineers for our courses to prepare themselves for the transition to IPv6.”

Cyrus Daruwala feels that the responsibility must be shared between the service provider and the company. Both parties must work closely together to ensure the Internet community can continue to access content on the company’s servers. “The content provider has to work with its Internet service provider to make sure its content is disseminated to an IPv6 capable server. You can move that system to the cloud to hasten the process.”

Tony Hill notes that some companies may find network translation solutions the viable temporary alternatives to implementing full scale IPv6 transition on their networks.“The scale of reworking can be huge. However, there are short circuit approaches using translation to make IPv4 sites visible to IPv6. Such an approach is viable only for smaller scale operations.”

Yew Hock Meng also notes that the Internet community is larger than just service providers and their customers, and sometimes connections may be outside of their control. “The responsibility to ensure downstream access via IPv4 and IPv6 is a collaborative effort between customers and providers. For customers with huge networks and high demand for IPv4, we have to work hand in hand. There are many issues, but no guarantees as providers are also at the mercy of their upstream IPv6 transit providers.”

Miwa Fujii agrees that IPv6 is not immediately critical, but she sees it as a case of the earlier the better for business reasons. “In the future, there will only IPv6 addresses available. If your resources are not IPv6 capable, you may lose cash flow, because new companies will only be given IPv6.”

Cyrus Daruwala also supports early transition as it will eventually lead to lower total costs of ownership. “The need for interoperability will compel you. If you can only respond to IPv4 and need to convert IPv6 traffic to IPv4, there will be latency. The conversion cost is also not cheap, and sooner or later, IPv6 will become a demand. It is not a do or die situation, but customer experience, interoperability and demand will be issue. Then it will be an escalating cost.”

Tony Hill’s perspective is that companies tend to procrastinate if they think it is not urgent, and eventually, may react later than they expect to. “I don’t think we have a long time. Our experience is that the world reacts better with a date. So, I created 8 Jun 2011 as IPv6 day this year. It gave a focal point for technical, media, and everyone else. It proved effective in garnering support. It meant that no single provider had to travel the road alone. There will be a date for world IPv6 week next year. The end of 2012 is a significant timing point for most economies around the world. They will be doing IPv6 by that time. That is not a bad planning horizon time.”

Yew Hock Meng adds that companies who have ample supplies of IPv4 addresses have an advantage, but still need to take action. “With your reserve of IPv4 addresses, you are fortunate. You don’t have to do translation. Take a very simple approach. Make your network dual stack. That is good enough. By the time you exhaust your IPv4 addresses, IPv6 will be dominant.”

Cyrus Daruwala notes that the key benefit is about sustained growth, which only viable in the Internet world with IPv6. “If you can’t cater to the consumer, it goes against business continuity. If you can scale up, then it also feeds into growth. From this perspective, the drive is not what the government is doing. It is the consumer that will drive it; he wants instant, connectivity, security. That can only be on IPv6.”

While most companies think about IPv6 as added costs, Tony Hill takes a different tack and feels that the potential of IPv6 has been vastly underrated. “I think the feel that there is no money to be made in IPv6 is slightly misunderstood and misleading. The analysis is very one sided. For ISPs, the analysis is all about cost. The other side of revenue, growth, and benefits should be considered. Very few analyse both sides and implement that. In fact, very few people are looking at the detailed list of technical benefits. There are huge business benefits available.”

Yew Hock Meng, agrees that IPv6 is potentially highly profitable, but more importantly, it provides significant competitive advantage. “I believe we should look at adoption of IPv6 as a measure of economic competitiveness. The moment you are connected, you will have a very strong impact on economic growth and interactivity around the world. It is important to be connected. How can you grow if you don’t have scalability? I can’t give more on IPv4. The only way forward is IPv6. While we see tons of money to be made, we have to see continuity. It is not a service provider or content issue, it is a national issue.”

Willy Sutrisno, ended the discussion by noting that while moving to IPv6 is imperative, it is even more critical for Singapore and Asia. “Singapore falls under APNIC for IP allocation. It is inevitable that IPv4 resources are finished in Asia first. Even if you have reserve IPv4 allocations, what about others? You have to reach the IPv6-only people.”