Prologue

1
(No Transcript)
2
Prologue
Self-Citation More than 7 papers at places of
least relevance
Nothing new except for the problem
We stress however that our proposal is somewhat
motivated by Our proposal, along the lines
described in 33
- Taken right from the paper
Meet the paper without Conclusion ? (Ok There is
a section on Discussion)
3
Towards an Internet Architecture
Evolvable
Authors Sylvia Ratnasamy Scott Shenker Steven
McCanne Presentation By Rahul
Potharaju Theodore Brockly
Something, somewhere, went terrible wrong!
4
To prove 21, well, assume that 12!
Even Asimovs Robotics laws will come to their
knees! Assumptions spoilt the game
For instance, take A4 ? Revenue Flow We know
that IPvN is great because of its benefits but
who will convince the normal not-so-hi-tech
consumers? Alright, ISP will claim Faster,
Faster and even more Faster ? cliché
5
Behind the Scenes
So let us consider IPv8 but wait What does it
contain in the first place?
What if the long awaited IPv8 increases
reliability and security at the cost of
throughput? No more advertisements for the ISP!
IPv8 ? Encapsulation of all IPv8 packets through
non IPv8 networks ? Even lower
throughput ? What will happen to the
users who chose the advanced ISP?
What if the revenue gained is offset due to the
overhead incurred by the initial implementation
of the new architecture?
6
Is it New? Different?

• Anycast is proposed as an alternative to ISP
  upgrades and Overlays.
• Two methods for deployment of Anycast
• ISP Policy Change
• Aggregatable address Overlay
• Anycast is simply a lightweight routing Overlay

7
Or Any Better?

• Anycast still requires several next gen service
  providers to show any gains.
• Will Anycast scale? ? They say they are not
  concerned about scalability!
• Completely mitigate the gain of new
  architectures?
• No one knows.
• Would still require an extraordinary amount of
  cooperation between ISPs.

8
Economic Concerns

• Running Anycast discourages evolution just as
  much as overlays
• First mover advantage almost nonexistent
• Multiple adopters needed to show gains
• No incentive whatsoever to deploy.
• Why should I help out Next Gen competitor X?

9
Capability ! Evolution

• Even if the architecture was adopted against
  ISP's interests
• Opens the door for a variety of incompatible
  protocols
• Across networks, nations
• Creates hidden costs through added complexity
• Complicates current work in the IP space (such as
  capability systems)?
• Does not support router dependent protocols

Unfortunately our approach does not assist in the
deployment of architectures certain QoS
proposals
- Taken from the paper
10
Possibility of Abuse War of the ISPs
IPv8
Nice that I am getting money from settlement
payments
Some or all of them will block IPv8
Why am I subsidizing X?
So what should the customers do? Spend time and
money to switch providers Simply return to IPv4
(more likely because none of us even know what
IPv8 is!)
ISP X
ISP A
ISP C
ISP B
11
Whats our point?
ISPs are not simply going to let forerunners
steal all of their business, and may not simply
surrender and upgrade their IP architectures to
the highest level due to the overhead of making
changes.
12
Anycast for Anything?
To use IPv8, any end host can simply encapsulate
an IPv8 packet in an IPv4 packet with destination
A4
- Taken right from the paper
How do you rule out the possibility of spoofing,
ARP poisoning etc?
Biggest issue with Anycast itself ? Any service
which relies on it might not function properly in
the case that the address is blackholed due to
whatever reasons, or if anyone close by isnt
offering service, you could end up traveling to
the ends of the earth to get DNS service.
13
Final Points
Good Intentions but bad reasoning throughout the
paper After reading your paper, we thought
that perhaps a revolution is needed just to
understand the contribution! More elaboration
on how to fix the architectural flaw of the lack
of incentives, but not a mere walk-through of
design options and their implications with little
to offer in novel mechanisms or radical insights
14
Back to the Defense Team
15
(No Transcript)
16
(No Transcript)
17
From a creators POV NGSP?
18
Possibility of Abuse
Let us say IPv8 becomes famous and everyone wants
it
The brilliant ISP X has implemented it
? All IPv8 traffic from ISPs A,B,C etc will be
going through ISP X provided that IP Anycast
routing and advertising-by-proxy is being used.
But these rival ISPs are not so happy they are
subsidizing their biggest competitor, who,
although not drawing large number of new
subscribers, is getting money through settlement
payments due to the traffic
? Some or all of them will start blocking the
IPv8 traffic!
So what do the customers do? Spend time and
money to switch providers Simply return to IPv4
(more likely because none of us even know what
IPv8 is!)
19
An unparalleled claim?
While many discuss the incremental deployment of
their solutions, they address the technical but
not economic issues of partial deployment we
look for what is missing from todays
architecture that would make evolution by its
incumbent operations economically desirable.
- Taken right from the paper
What is the new architecture? What was addressed
was How the new architecture and why it will
propagate and be successful!