Network and Telecom Strategies Blog

Time to inventory your IPv4 Addresses!

Posted by: Eric Siegel

I see that Mike put up a blog post about IPv6 ("And you think 2012 will be a problem...") mentioning that I'm working on a paper about IPv4 address exhaustion and the resulting need to be prepared for IPv6-only Internet users.

An important consideration for all enterprises is that, starting in 2009, some Regional Internet Registries (RIRs) began to formalize mechanisms for the transfer of address blocks subject to RIR approval. That allows the appearance of a market in address blocks.

The address block market will peak as the IPv4 address shortage becomes extreme after 2012, then gradually decline as IPv6 gains popularity. Large blocks of addresses, which had no financial value previously, will need to be inventoried and appear as an enterprise asset. (Larger blocks will be more valuable than an equivalent number of addresses in smaller blocks.) In some cases, enterprises may re-number their internal networks to create large address blocks that can then be sold.

So, if you have no idea how many IPv4 address blocks you own, or which ones are used, the time has arrived for you to start an inventory. IPv4 address blocks will suddenly have financial value, and you might even want to consider selling them in a couple of years to an enterprise that hasn't planned ahead and suddenly needs IPv4 addresses but can't get them from their ISP. (That won't be YOU, of course....)

By the way, you can talk with Mike in person about IPv6 and IPv4 at the upcoming Catalyst Conference in Europe on April 19-22. He'll be there!

If you're interested, here are some of the RIR regulations governing transfer of IPv4 blocks:

RIPE:

http://www.ripe.net/ripe/docs/ripe-484.html#55

ARIN:

https://www.arin.net/resources/request/transfers.html ("Transfer of released number resources")

http://lists.arin.net/pipermail/arin-ppml/attachments/20090406/b8a75e31/attachment.pdf

APNIC:

http://www.apnic.net/policy/proposals/prop-050

The Blivet and IPv6

Posted by Michael Disabato

When I was in engineering school and working as an intern, I was introduced to the term "blivet." According to Wikipedia (http://en.wikipedia.org/wiki/Blivit), "In traditional U.S. Army slang dating back to the Second World War, a blivet was defined as "ten pounds of manure in a five pound bag" (a proverbial description of anything egregiously ugly or unmanageable); it was applied to an unmanageable situation, a crucial but substandard or damaged tool, or a self-important person." What does this have to do with networking and IPv6? Everything.

I have recently held discussions with several governmental organizations, retail, medical, and financial firms about IPv6 and whether or not they should proceed with it. In every case there has been interest, but no movement, mostly citing lack of business case. Well, I have one for you.

Several of these firms are exhausting the private IP space they are using. One of them had this happen overnight as they absorbed one of their former competitors, which had a network almost as big as theirs. One firm is close to hitting limits because they are virtualizing their desktops, and I don't mean thin clients accessing instances of Windows on a central machine. I mean multiple instances of multiple operating systems on their workstations. Each of these requires an IP address, so you can see where this is going. We may be looking at public IPv4 address exhaustion in one or two years, but some firms see it now. They have a 5 pound network with 10 pounds of addresses. Blivet.

One practice to alleviate this problem is using "dark" IP addresses. Basically you take a range of addresses assigned to another registry and treat them as unroutable, hiding them behind your firewall. This is all well and good until you either acquire the company that really owns the addresses and have to deal with overlapping address space or Something Goes Wrong with your firewall and this hidden address space becomes visible (oops). The latter can bring all sorts of collateral damage when news of this hits the press.

Another practice is to segment the network and create NAT areas. However, this can impact legacy applications, real-time applications (voice and video), and peer-to-peer applications. 

What's the solution to this? Get a bigger bag. There is no longer time to wait on investigating and certifying IPv6 for use in the enterprise. Either a merger, acquisition, or virtualization of your infrastructure will force you to this before we run out of IPv4 addresses. The recommendations from Burton Group are:

• Set up a lab and begin testing and learning immediately 
• Enable your edge connection to the Internet to handle IPv4 and IPv6 
• Identify applications and infrastructure components that cannot be upgraded to hand IPv6 and make plans to isolate them in IPv4 islands with appropriate gateways.
• Set up dual stacks (IPv4 and IPv6) into your data center for you internal network
• Inventory your IP address space and keep a close watch on consumption rates
• Make sure all new applications can handle IPv4 and IPv6 and retrofit existing applications during maintenance cycles 

Burton Group has documents that cover this in detail, and our consulting group can help you plan for the migration and adoption of IPv6. Tempus fugit, folks. 

Michael 

'Tis the season

Another holiday season has come and gone and if you're like me, although holidays are supposed to be relaxing they're always a bit too hectic to be doing any serious planning - or making meaningful resolutions.  As a rule, it's important to take stock of the past years successes and failures and now that the kids are back in school and the work routine has begun again it's time to review 2008 and think about 2009.  But do it quick before the work routine buries you again. 

The top ten (or top 21, 50, 100...) lists are a good place to start and fortunately all of the media outlets publish some form of "Best/Worst _____ of 2008."   They often publish a shorter list of predictions for the coming year too.  After the long winter's nap the right lists can really spark your memory of 2008 and get creative juices flowing again for 2009.  When you're reading the lists decide what you agree with and what doesn't add up, perhaps even what's missing and should go into your business planning or your personal resolutions.

You may have already read 100 emails touting one or more of the articles I've mentioned below, but check out a few and use them to do your own planning.  

Fortune's 21 dumbest moments in business

Time's Best Inventions of 2008

or

Time's 50 Best Web Sites of 2008

need a new laptop? see:

PC World's Best Laptops of 2008

enjoy other people's misery? see:

CIO magazine's top ten email blunders of 2008

interested in the best and worst attempts to regulate the Internet? see:

InformIT

Maybe a summary of 2008 in the form of awards from one of our favorite tech blogs, ArsTechnica

or a best of the best from Lifehacker's Most Popular Top 10's of 2008. 

One of the year-end articles that caught my attention came from ArsTechnica and concerns the exhaustion of IPv4 address space on the Internet (see IPv4 Addresses: Almost 200 million served in 2008).  In 2008 I wrote a report on this topic for the Burton Group and this ArsTechnica post reminded me to continue to reach out and urge the IT world to start thinking about IPv6.  

Although many large enterprises may have all of the IPv4 space they need, the bottom line is this:  the time when large service providers can no longer find large blocks of IPv4 address space to use is fast approaching.  The most likely place for them to turn?  IPv6.  This means IPv6-only hosts (those that only speak IPv6) are going to appear on the Internet -- is your enterprise ready to speak to them?

IPv6 - Ready or not, here it is

Posted by: Michael Disabato

John Curran, Chair of ARIN (American Registry of Internet Numbers), spoke at Catalyst today on the depletion rate of IPv4 addresses. By itself, this could have been viewed as another "doom and gloom" prediction similar to the one's we've heard in the past. However, when placed in context of previous predictions, this one is enough to make me think it's time to take IPv6 out of the lab and start testing in live environments. (What? You're not testing IPv6 in a lab environment?)

Two years ago, the prediction was depletion of IPv4 addresses in 2025. Last year, the prediction was 2017. This year, John showed a chart claiming we'd be out of addresses in 2012. If that trend continues, next year will see the curve move to 2009. Regardless, depletion of IPv4 addresses means IPv6 will be forced on new companies attempting to connect to the Internet. While your network will continue to run, there is the distinct possibility you will not be able to connect to companies running only IPv6 without an IPv6 connection on your network. This, of course, will be a BAD THING. (Those of you dealing with DoD have been living with this possibility for a while.)

Burton Group has changed its stance of two years ago from "IPv6 won't be needed for a while" to last year's "Get familiar with it by testing in a lab." Based on this year's prediction, perhaps it's time to start limited deployments of IPv6. It's in your equipment. It's in the desktop and server operating systems. Turn it on and play with it. Get familiar with it. See how to break it, and most importantly, see what it breaks. And *this* time, learn how to manage it before you deploy it.

This is a non-trivial project, so the sooner you start, the sooner you'll be ready.