By now most of the IT industry has finally realized that IPv6 is soon becoming inevitable. But many people don't realize what will happen with IPv4 some time throughout this year: As soon as ISPs introduce Carrier Grade NAT (CGN), the resulting functional limitations won't just cause some extra work in the IT departments but will render popular business models infeasible.
The talk gives an overview what to expect during the upcoming final stage of IPv4 and the introduction of IPv6, how to handle it and why open source is so crucial during this period.
Architecture decision records - How not to get lost in the past
OSDC 2012 | Like rats on a sinking ship by Benedikt Stockebrand
1. Like Rats on a Sinking Ship
Benedikt Stockebrand
Open Source Data Center Conference
April 2012
N¨urnberg (Nuremberg, Germany)
Copyright c 2012 Benedikt Stockebrand 1/18
2. About Me
• Diplom-Informatiker (Uni Dortmund)
• Internet since approx. 1993
• Focus on IT operations, TCP/IP networks, Unix,
(some aspects of) Security
• 2000–mid 2002 employee of T-Online, Data Center
Management Team
• Since 2003 freelance trainer and consultant,
with IPv6 main area of interest
• Author of “IPv6 in Practice—A Unixer’s Guide to the Next
Generation Internet” (Springer 2006/2007)
• Co-author of a study for the German federal office for IT
security/BSI (2010, with Hans Peter Dittler)
• Currently starting a training and consulting company with
focus IPv6
Copyright c 2012 Benedikt Stockebrand 2/18
3. IPv6 and the TCP/IP Stack I
Application
Layer
Transport
Layer
Network
Layer
Link
Layer
DNS SSH SMTP IMAP HTTP ···
TCP UDP ···
IP(v4)
IGMP ICMP
IPv6
MLD ICMP6
Ethernet PPP WLAN ···
Copyright c 2012 Benedikt Stockebrand 3/18
4. IPv6 and the TCP/IP Stack II
Physical+Link Layer
Application Layer
Transport Layer
Network Layer
Copyright c 2012 Benedikt Stockebrand 4/18
5. IPv6 and the TCP/IP Stack III
Copyright c 2012 Benedikt Stockebrand 5/18
6. IPv6 and the TCP/IP Stack III
Users, Developers, Admins
Copyright c 2012 Benedikt Stockebrand 5/18
7. IPv6 and the TCP/IP Stack III
Users, Developers, Admins
(Non-technical) Management
Copyright c 2012 Benedikt Stockebrand 5/18
8. IPv6 and the TCP/IP Stack III
Users, Developers, Admins
(Non-technical) Management
Politics
Copyright c 2012 Benedikt Stockebrand 5/18
9. IPv6 and the TCP/IP Stack III
Users, Developers, Admins
(Non-technical) Management
Politics
Churches
Copyright c 2012 Benedikt Stockebrand 5/18
10. Is Vintage IPv4 a dying business model?
Yes,
• new business ideas fail due to lack of addresses
• keeping Vintage IP alive gets increasingly expensive
• functionality will be reduced
• . . . like reliability
• business models based on GeoIP location will break down
Copyright c 2012 Benedikt Stockebrand 6/18
11. Is Vintage IPv4 a dying business model?
No,
• large scale Carrier Grade NAT (CGN) becomes unavoidable
• . . . and hardware vendors will happily sell the gear needed
• . . . and even throw in some “transition technologies”
(for not so free)
• . . . and workarounds for CGN limitations become a product
Copyright c 2012 Benedikt Stockebrand 7/18
12. Is Vintage IPv4 a dying business model?
No,
• large scale Carrier Grade NAT (CGN) becomes unavoidable
• . . . and hardware vendors will happily sell the gear needed
• . . . and even throw in some “transition technologies”
(for not so free)
• . . . and workarounds for CGN limitations become a product
• Address “relocation” becomes a serious business model
• . . . at everybodies expense
Copyright c 2012 Benedikt Stockebrand 7/18
13. What’s the Business Value of IPv6?
Copyright c 2012 Benedikt Stockebrand 8/18
14. What’s the Business Value of IPv6?
• None
Copyright c 2012 Benedikt Stockebrand 8/18
15. What’s the Business Value of IPv6?
• None
• . . . but. . .
Copyright c 2012 Benedikt Stockebrand 8/18
16. What’s the Business Value of IPv6?
• None
• . . . but. . .
• Y2K didn’t have a business model either
Copyright c 2012 Benedikt Stockebrand 8/18
17. What’s the Business Value of IPv6?
• None
• . . . but. . .
• Y2K didn’t have a business model either
• neither had that entire Internet Thing[TM]
Copyright c 2012 Benedikt Stockebrand 8/18
18. What’s the Business Value of IPv6?
• None
• . . . but. . .
• Y2K didn’t have a business model either
• neither had that entire Internet Thing[TM]
• the business model is to get rid of Vintage IPv4
• in business language:
“Not having IPv6 incurs extensive opportunity costs”
Copyright c 2012 Benedikt Stockebrand 8/18
19. What Happens Now?
• Vintage IPv4 addresses should last until August in RIPEland
• . . . unless there’s a panic before that (like in APNICland)
• Then Vintage IPv4 will become comatose
• . . . and we’ll do significant intensive-care work
• IPv6 should then deploy explosively
• . . . and whoever waited too long will lose market share
Copyright c 2012 Benedikt Stockebrand 9/18
20. ISP Related Problems
• Everyone waits for them
• Can’t use Vintage IPv4 as a fallback
• Must decide on the access technologies
• Need significant networking skills
• Take a heavy hit on their first level support
Copyright c 2012 Benedikt Stockebrand 10/18
21. Customer Premises Equipment
• Implementing all transition technologies is expensive
• Any “do not press” button you add will be pressed. . .
Copyright c 2012 Benedikt Stockebrand 11/18
22. Network Hardware, Enterprise Class
• Timing of IPv6-capable products critical
• IPv6 support still sometimes limited
• Long product lifetime/depreciation time
• Latest moment features troublesome
Copyright c 2012 Benedikt Stockebrand 12/18
23. Software
• OSes are ok
• Standard software is largely ok
• Specialized software is particularly troublesome
• Open source has a significant headstart
• Some public funding was given (WIDE/KAME project)
• Developers don’t have to convince management
• Comatose software will die
• . . . and we’ll have to switch products
Copyright c 2012 Benedikt Stockebrand 13/18
24. The Support Hierarchy
End Users
“Power Users”/”Zero Level Support”
Help Desk/First Level Support
System/Network Administration
DevelopmentEscalation
Copyright c 2012 Benedikt Stockebrand 14/18
25. People Problems
• Tech crowd: Old dogs need to learn new tricks. . .
. . . from new dogs
• (Ex-)technical management: Need to let go
• “Real” (non-technical) management:
Have to face an unfamiliar reality
Copyright c 2012 Benedikt Stockebrand 15/18
26. The Great Clean-Up
• IPv6 will bring up all the legacy problems collected since Y2K
• Everyone will be affected, but
• Starting early saves money
• Having kept your IT environment clean saves money
• Flexibility to fix things fast gains market share
• IPv6 is a chance to straighten things out
Copyright c 2012 Benedikt Stockebrand 16/18
27. Teething Troubles
• Security
• The privacy discussion
• Getting static addresses
• ISPs and flatrates
Copyright c 2012 Benedikt Stockebrand 17/18