MiSMoSS

1
MiSMoSS
Middleware Services for Management of Shared
Space in Large-Scale Distributed Interactive
Applications

• IFHS V05
• 5/1 2005

2
Evolution RequirementsTextual commands and
textual chat
3
Evolution RequirementsAV chat and AV
conferencing
4
Evolution RequirementsHaptic Interaction
5
Evolution RequirementsAll requirements in one
application
6
MiSMoSS System Support for Large-Scale
Interactive Applications

• Application area interactive distributed
  multimedia applications
• Continuous and non-continuous, interactive and
  non-interactive media
• Event streams
• Audio and video
• Large-scale
• Long distances
• Many concurrent users
• Goal provide support for the best possible
  perception to users
• Shield users from some effects of resource
  limitations
• Overcome challenges with large geographical
  distances

7
MiSMoSS
Real-World Proximity
Virtual World Proximity
Reference applicationLarge-Scale, Interactive,
Distributed Games
8
MiSMoSS
Real-World Proximity
Virtual World Proximity


Large numberof users
Possiblepriorization
Long distances
Limited userperception
Computer-controlledparticipants
Interestmanagement
9
MiSMoSS
10
MiSMoSS
Sub-topic 1 Latency Hiding
(prediction, time synchronization, )
11
MiSMoSS
12
MiSMoSS
Sub-topic 2 Data Reduction
(region-of-interest, content-based routing,
filtering, multicasting, )
13
MiSMoSS
Sub-topic 3 Data replication (proxy
caching servers, multicasting, fault tolerance,
consistency, )
14
For more information

• Members / persons
• Chris Majewski (room 3216B, email krzys _at_ ifi)
• Knut-Helge Vik (room 3216B, email knuthelv _at_
  ifi)
• Wladimir Palant (room 3216B, email palant _at_
  ifi)
• Carsten Griwodz (room 3343, email griff _at_ ifi)
• Pål Halvorsen (room 3345, email paalh _at_ ifi)
• WWW
• http//www.ifi.uio.no/forskning/grupper/nd/project
  s/2004/mismoss/index.html

15
IXP Internet Exchange Processor
Use Network Processors to Increase Performance in
a Distributed Environment

• IFHS V05
• 5/1 2005

16
Forwarding Example for an Intermediate Node
Intel Hub Architecture
application
user space kernel space
Note- one single average MPEG-II DVD stream
require 330-660 packets per second of 1500
Bytes (4-8 Mbps) - then use smaller packets, add
concurrent clients, other applications,
communication system
Pentium 4 Processor
registers
cache(s)
communication system
application
network card
17
Main Packet Processing Costs

• Copying used when moving a packet from one
  memory location to another
• expensive (proportional to packet size)
• should be avoided whenever possible (use
  pointers)
• Checksuming used to detect errors
• expensive (proportional to packet size)
• transport layer payload header
• network layer header
• Fragmentation/reassembly needed when packet is
  larger than smallest MTU
• generate headers header checksum
• receiving many small data fragments

18
Review of Conventional Computer Hardware
Architectures
Intel D850MD Motherboard - Intel Hub Architecture
(850 Chipset)
RDRAM connectors
CPU socket
RDRAM interface
system bus
hub interface
PCI bus
Memory Controller Hub
I/O Controller Hub
PCI connectors
19
Network Processors Main Idea
Traditional system - slow - resource demanding -
shared with other operations
IXP network processor - a computer within the
computer - special, programmable hardware -
offloads host resources
20
IXA Internet Exchange Architecture

• IXA is a broad term to describe the Intel network
  architecture (HW SW, control- data plane)
• IXP Internet Exchange Processor
• processor that implements IXA
• IXP1200 is the first IXP chip (4 versions)
• IXP2xxx has now replaced the first version

• IXP1200 basic features
• 1 embedded 232 MHz StrongARM
• 6 packet 232 MHz µengines
• onboard memory
• 4 x 100 Mbps ethernet ports
• IXP2400 basic features
• 1 embedded 600 MHz XScale
• 8 packet 600 MHz µengines
• 3 x 1 Gbps optic Ethernet ports

21
IXP results so far way ahead

• Can do a lot on microengines StrongARM
• Forwarding is extremely fast compared to sending
  the packet to the host good for interactive
  applications
• Some examples are
• (Proxy) Server topics, including header
  processing/header templates, reporting, routing
  RTP packets, adding FEC, TCP-friendly schemes,
  TCP congestion prediction, and interrupt
  collection/reduction, etc.
• Reflector/filter/router topics, including
  filtering by packet priority/congestion
  information, priority queues, format conversion,
  packet reordering and dropping, reservation,
  jitter compensation, shaping, etc.
• Combine on-board processing and OS enhancements
• Accelerated group communication in interactive
  distributed applications such as games
• Experiments for ICON / VINE

22
For more information

• Persons
• Carsten Griwodz (room 3343, email griff _at_ ifi)
• Pål Halvorsen (room 3345, email paalh _at_
  ifi)
• WWW
• http//www.ifi.uio.no/griff/theses/course18.html

23
Media Streaming
Other topics related to media streaming,
including adaptation to varying network
conditions, kernel extensions and possibly use
of network processor cards

• IFHS V05
• 5/1 2005

24
Delivery Systems
Network
25
Delivery Systems
Users selection
Network
Saving network resources Stream scheduling
26
Delivery Systems
bus(es)
27
Delivery Systems
CPU scheduling
File system
Protocols
Application buffer mgnt
Kernel buffer mgnt
Diskscheduling
Network I/O scheduling
Disk layout
bus(es)
28
Delivery Systems
Same OS support
Network
Network
Proxy server
Central server
Proxy placement and dimensioning
Data selection and upload
29
Distributed Applications
30
Distributed Applications

• Network processor cards
• Multimedia file systems
• Zero-copy memory architectures
• Memory caching
• TCP-friendly protocols
• Multimedia servers
• Resource management

31
Distributed Applications

• Content distribution networks
• Caching and replication
• Guaranteed and adaptive QoS
• Resource management

32
For more information

• Persons
• Carsten Griwodz (room 3343, email griff _at_ ifi)
• Pål Halvorsen (room 3345, email paalh _at_
  ifi)
• WWW with example topics
• http//www.ifi.uio.no/griff/theses/theses.html