Intro to Microkernel Study Group

1
Intro to Microkernel Study Group
2
What its about

• Microkernels and microkernel construction
• Building systems upon microkernels
• Research issues

3
Background the bigger picture

• What are directions and goals of the group?
• What do we hope to achieve?

4
Vision

• a new Paradigm
• i.e. an orthogonal set of concepts
• that can be implemented extremely efficient
• for any hardware
• for any type of application,
• a basis for arbitrary systems
• enabling fine-granular component-based system
  construction

5
General Goal

• define and establish microkernel technology

6
Strategic Goals

• Manage system complexity
• Weaken legacy dependance
• conceptual-basis for system architecture

7
Tactical Goals

• Wide Applicability
• For all hardware architectures
• For all system and application architectures
• Open Backbone Project
• Client Projects
• Simple Access and Participation
• Streamlining Maintenance

8
Technical Goals

• Evaluating Limits of Efficiency
• Standardprocessors x86
• Low-Power ARM
• 64-bit IA-64, Power4
• Massively-parallel Architectures
• Maximal Flexibility and Generality
• Conventional Systems
• Pervasive Systems
• Superservers
• SASOS
• Deep computing, ...

9
What are we doing more specifically

• L4Ka project
• SawMill prime project

10
Ver 4
Orange Pip
Superfast IPC
Pistachio
SMP mechanisms
Thread naming
Non monotonic clocks
Denial- of-service attacks
Time abstr.
Cashew
Grapestone
Delayed preemp- tions
Persistence
Events
Covert channels
IA-64 registers
RT-sched mechanisms
SASOS support
Kernel compiler
Threads as virtual objects
Huge memories
Memo- ry-bus schedul.
Large multi-level caches
Numa mechanisms
Compress memory
Power control
Ver 5
Fast intercon- nects
Multi- threaded proc.
Fault-tol. support
Multigiga- bit cluster nets
11
SawMill Prime
Stub Generation IDL4
Component Composition
Version Build Managment
Buffer Management
Component Specification
Orthogonal Components
Resource Management
Naming
VM Framework
Driver Framework
Buffer Transfer
Memory Pinning
Legacy Component Extraction
Dataspace Stacking
Synchronization
Security
Performance
Performance
Performance
12
What to expect

• Presentations
• Diplomarbeits, Studiarbeits
• Research issues
• Discussion topics
• Active debate
• Heated discussions
• Flame wars
• The occasional quiet week.

13
Problem

• Being research, some topics might be beyond your
  background.
• We will try not to assume to much background.

14
How to get something out of this

• If you dont understand something,
• Ask questions!!!

15

• See you next week!!