WAMOS 2018: Thorsten Knoll thorsten.knoll@hs-rm.de Adapting - - PowerPoint PPT Presentation

wamos 2018 thorsten knoll
SMART_READER_LITE
LIVE PREVIEW

WAMOS 2018: Thorsten Knoll thorsten.knoll@hs-rm.de Adapting - - PowerPoint PPT Presentation

WAMOS 2018: Thorsten Knoll thorsten.knoll@hs-rm.de Adapting Kerckhoffs principle: A path from cryptography to open source hardware Adapting Kerckhoffs principle Content of presentation The path: The idea Kerckhoffs principle (KP)


slide-1
SLIDE 1

WAMOS 2018: Thorsten Knoll

thorsten.knoll@hs-rm.de

Adapting Kerckhoffs principle:

A path from cryptography to open source hardware

slide-2
SLIDE 2

Adapting Kerckhoffs principle WAMOS 2018, Thorsten Knoll Slide 2 / 31

The path:

  • The idea
  • Kerckhoffs principle (KP)
  • Attacks and KP
  • Mitigations and KP
  • The need for Open Source Hardware (OSH)
  • Conclusion: Adapting KP

Content of presentation

slide-3
SLIDE 3

Adapting Kerckhoffs principle WAMOS 2018, Thorsten Knoll Slide 3 / 31 The idea

The idea

slide-4
SLIDE 4

Adapting Kerckhoffs principle WAMOS 2018, Thorsten Knoll Slide 4 / 31 The idea

New: Sidechannel attacks on mainstream CPUs (~2000) Brandnew: Meltdown, Spectre, BranchScope, TLBleed, Spectre-NG, … (2018) Very old: Kerckhoffs principle (1883)

slide-5
SLIDE 5

Adapting Kerckhoffs principle WAMOS 2018, Thorsten Knoll Slide 5 / 31 The idea

New: Sidechannel attacks on mainstream CPUs (~2000) Brandnew: Meltdown, Spectre, BranchScope, TLBleed, Spectre-NG, … (2018) Very old: Kerckhoffs principle (1883) Paper: Adapting Kerckhoffs principle

+ +

slide-6
SLIDE 6

Adapting Kerckhoffs principle WAMOS 2018, Thorsten Knoll Slide 6 / 31 Kerckhoffs principle (KP)

Kerckhoffs principle (KP)

slide-7
SLIDE 7

Adapting Kerckhoffs principle WAMOS 2018, Thorsten Knoll Slide 7 / 31 Kerckhoffs principle (KP)

1883 - Auguste Kerckhoffs in “La cryptographie militaire”: Design principles for (military) cipher systems:

  • 1. The system must be practically, if not mathematically, indecipherable.
  • 2. It should not require secrecy, and it should not be a problem if it falls

into enemy hands.

  • 3. It must be possible to communicate and remember the key without

using written notes, and correspondents must be able to change or modify it at will.

  • 4. It must be applicable to telegraph communications.
  • 5. It must be portable, and should not require several persons to handle
  • r operate.
  • 6. Lastly, given the circumstances in which it is to be used, the system

must be easy to use and should not be stressful to use or require its users to know and comply with a long list of rules.

Translation from french to english: https://en.wikipedia.org/wiki/Kerckhoffs%27s_principle

slide-8
SLIDE 8

Adapting Kerckhoffs principle WAMOS 2018, Thorsten Knoll Slide 8 / 31 Kerckhoffs principle (KP)

1883 - Auguste Kerckhoffs in “La cryptographie militaire”: Design principles for (military) cipher systems:

  • 1. The system must be practically, if not mathematically, indecipherable.
  • 2. It should not require secrecy, and it should not be a problem if it falls

into enemy hands.

  • 3. It must be possible to communicate and remember the key without

using written notes, and correspondents must be able to change or modify it at will.

  • 4. It must be applicable to telegraph communications.
  • 5. It must be portable, and should not require several persons to handle
  • r operate.
  • 6. Lastly, given the circumstances in which it is to be used, the system

must be easy to use and should not be stressful to use or require its users to know and comply with a long list of rules.

Translation from french to english: https://en.wikipedia.org/wiki/Kerckhoffs%27s_principle

slide-9
SLIDE 9

Adapting Kerckhoffs principle WAMOS 2018, Thorsten Knoll Slide 9 / 31 Kerckhoffs principle (KP)

1883 - Auguste Kerckhoffs in “La cryptographie militaire”: Design principles for (military) cipher systems:

  • 2. It should not require secrecy, and it should not be a problem if it falls

into enemy hands. Broader interpretation: A cryptosystem’s security must solely depend on keeping the keys secret, not the algorithms. Now known as “Kerckhoffs principle” (KP)

slide-10
SLIDE 10

Adapting Kerckhoffs principle WAMOS 2018, Thorsten Knoll Slide 10 / 31 Kerckhoffs principle (KP)

KEY

Cryptosystem (Algorithms) Plaintext (Data) Plaintext (Data) Secret Open

slide-11
SLIDE 11

Adapting Kerckhoffs principle WAMOS 2018, Thorsten Knoll Slide 11 / 31 Kerckhoffs principle (KP)

1949 - Claude E. Shannon:

Reformulation of KP: The enemy knows the system. Design principle: A cryptosystem should be designed as if the enemy would already know the system. Now known as “Shannon’s maxim” (SM)

slide-12
SLIDE 12

Adapting Kerckhoffs principle WAMOS 2018, Thorsten Knoll Slide 12 / 31 Kerckhoffs principle (KP)

2002 - Bruce Schneier:

  • Generalization for all secure systems.
  • More secrets to be kept → More fragile.
  • Less secrets to be kept → More robust.
  • If algorithms, protocols, implementations have to be kept secret

→ They’re part of the (composite) key.

  • How easily and costly is the replacing of compromised keyparts?
  • Openness enables reviews and evaluation, implied a community.
  • Design principle:

Minimize the number of secrets in your security system.

slide-13
SLIDE 13

Adapting Kerckhoffs principle WAMOS 2018, Thorsten Knoll Slide 13 / 31 Kerckhoffs principle (KP)

2018 – Thorsten Knoll: KP Point Secret (Composite key) Open more fragile more robust Secure System

slide-14
SLIDE 14

Adapting Kerckhoffs principle WAMOS 2018, Thorsten Knoll Slide 14 / 31 Attacks and KP

Attacks and KP

slide-15
SLIDE 15

Adapting Kerckhoffs principle WAMOS 2018, Thorsten Knoll Slide 15 / 31 Attacks and KP

Since 2003:

  • Remote attacks are practical
  • Prime and Probe
  • Predicting secret keys via branch prediction
  • Cache games
  • Flush and reload
  • Evict and reload

Since 2018:

2018 Jan Feb Mar Apr May Jun Jul

Meldown, Spectre V1+V2 Branch- Scope Spectre NG V3a + V4 TLBleed Spectre NG V3 Spectre NG V1.1

slide-16
SLIDE 16

Adapting Kerckhoffs principle WAMOS 2018, Thorsten Knoll Slide 16 / 31 Attacks and KP

Secret (Composite key) Open Actual CPUs Before 2018

slide-17
SLIDE 17

Adapting Kerckhoffs principle WAMOS 2018, Thorsten Knoll Slide 17 / 31 Attacks and KP

Now Secret (Composite key) Open Actual CPUs Before 2018

slide-18
SLIDE 18

Adapting Kerckhoffs principle WAMOS 2018, Thorsten Knoll Slide 18 / 31 Attacks and KP

Now Secret (Composite key) Open Actual CPUs Before 2018 Compromised keyparts

slide-19
SLIDE 19

Adapting Kerckhoffs principle WAMOS 2018, Thorsten Knoll Slide 19 / 31 Mitigations and KP

Mitigations and KP

slide-20
SLIDE 20

Adapting Kerckhoffs principle WAMOS 2018, Thorsten Knoll Slide 20 / 31 Mitigations and KP

Mitigation strategies:

  • In Software → Mostly open.
  • Seems like a hot-fix for a bigger problem (in hardware).
  • A lot of work in a small timeframe.
  • Billions of affected devices sold.
  • Not every attack vector is mitigatable in software.
  • Call for openness arises: New ISA’s including timings.
slide-21
SLIDE 21

Adapting Kerckhoffs principle WAMOS 2018, Thorsten Knoll Slide 21 / 31 Mitigations and KP

Now Secret (Composite key) Open Actual CPUs Before 2018 Compromised keyparts

Mitigations fill the gap Attacks move the KP-Point

slide-22
SLIDE 22

Adapting Kerckhoffs principle WAMOS 2018, Thorsten Knoll Slide 22 / 31 The need for Open Source Hardware (OSH)

The need for Open Source Hardware (OSH)

slide-23
SLIDE 23

Adapting Kerckhoffs principle WAMOS 2018, Thorsten Knoll Slide 23 / 31 The need for Open Source Hardware (OSH)

2018 – Sovereignty in IT:

  • 60 pages, describing the actual state.
  • Tons of examples: Hardware-Trojans, -Backdoors, -Killswitches.
  • Defines 13 actionpoints to start with.
  • Security can’t be added to hardware by software.
  • If software is secure, the attacks target a level deeper (hardware).
  • Design and production of silicon is closed by now and should be
  • pened to gain back sovereignty.
slide-24
SLIDE 24

Adapting Kerckhoffs principle WAMOS 2018, Thorsten Knoll Slide 24 / 31 The need for Open Source Hardware (OSH)

EDA Tools Open Source Software (OSS) Secret (Composite key) KP Point Open Pre- production Manu- facturing

Wanted

slide-25
SLIDE 25

Adapting Kerckhoffs principle WAMOS 2018, Thorsten Knoll Slide 25 / 31 Conclusion: Adapting Kerckhoffs principle

Conclusion: Adapting Kerckhoffs principle

slide-26
SLIDE 26

Adapting Kerckhoffs principle WAMOS 2018, Thorsten Knoll Slide 26 / 31 Conclusion: Adapting Kerckhoffs principle

Conclusions:

  • Attacks move the KP-Point by research, not by design.
  • Mitigations fill the gap to secure systems again.
  • Mitigations are necessary. Billions of devices sold.
  • Kerckhoffs principle is adaptable for measuring the situation.
  • More Openness doesn’t solve the actual situation, but in long term it

could relax the short timeframe situation.

  • A strong community is needed, for review and evaluation.
  • Open Source Hardware is on the rise (RISC-V).
slide-27
SLIDE 27

Adapting Kerckhoffs principle WAMOS 2018, Thorsten Knoll Slide 27 / 31 Conclusion: Adapting Kerckhoffs principle

Think ahead:

  • NIST has gone the way of finding new cryptography algorithm

standards through open, public competitions.

  • Maybe we’ll see such a competition for PBUs, TLBs and PHTs in near

future?

slide-28
SLIDE 28

Adapting Kerckhoffs principle WAMOS 2018, Thorsten Knoll Slide 28 / 31 Conclusion: Adapting Kerckhoffs principle

Think ahead:

  • NIST has gone the way of finding new cryptography algorithm

standards through open, public competitions.

  • Maybe we’ll see such a competition for PBUs, TLBs and PHTs in near

future?

  • Why would anyone build an i7 in Open Source? It might not even be

possible.

  • Instead putting the focus on massive parallelisation. Think about some

thousand RISC-V on a single die. That would even be possible with FPGA prototyping in academia.

slide-29
SLIDE 29

Adapting Kerckhoffs principle WAMOS 2018, Thorsten Knoll Slide 29 / 31 Conclusion: Adapting Kerckhoffs principle

The path:

  • All of this is a path to more robust, secure and open systems.
  • Going the presented path might take decades.
slide-30
SLIDE 30

Adapting Kerckhoffs principle WAMOS 2018, Thorsten Knoll Slide 30 / 31 Conclusion: Adapting Kerckhoffs principle

The path:

  • All of this is a path to more robust, secure and open systems.
  • Going the presented path might take decades.

But we’ll never arrive there, when we don’t start moving!

slide-31
SLIDE 31

Adapting Kerckhoffs principle WAMOS 2018, Thorsten Knoll Slide 31 / 31 Conclusion: Adapting Kerckhoffs principle

The path:

  • All of this is a path to more robust, secure and open systems.
  • Going the presented path might take decades.

But we’ll never arrive there, when we don’t start moving! Thank you.