J. K. Kim

1
Portland

• J. K. Kim

2
Trend of Data Center
By J. Nicholas Hoover ,  InformationWeek June 17,
2008 0400 AM
200 million Euro
http//gigaom.com/cloud/google-builds-megadata-cen
ter-in-finland/
http//www.datacenterknowledge.com

• Data centers will be larger and larger in the
  future cloud computing era to benefit from
  commodities of scale.
• Tens of thousand ? Hundreds of thousands in of
  servers
• Most important things in data center management
• Commodities of scale
• Keep High utilization of their expensive
  equipment
• Maximize revenue
• Saving administration cost
• Low power consumption (http//www.energystar.gov/
  ia/partners/prod_development/downloads/EPA_Datacen
  ter_Report_Congress_Final1.pdf)

3
Requirements for future data center

• To catch up with the trend of mega data center,
  DCN technology should meet the requirements as
  below
• High Scalability
• Transparent VM migration (high agility)
• Easy deployment requiring less human
  administration
• Efficient communication
• Loop free forwarding
• Fault Tolerant
• Current DCN technology cant meet the
  requirements.
• Layer 3 protocol can not support the transparent
  VM migration.
• Current Layer 2 protocol is not scalable due to
  the size of forwarding table and native
  broadcasting for address resolution.

4
Overall Idea of PortLand
Add a new host

• Key features
• Layer 2 protocol based on tree topology
• PMAC encode the position information
• Data forwarding proceeds based on PMAC
• Edge switchs responsible for mapping between
  PMAC and AMAC
• Fabric mangers responsible for address
  resolution
• Edge switch makes PMAC invisible to end host
• Each switch node can identify its position by
  itself
• Fabric manager keep information of overall
  topology. Corresponding to the fault, it
  notifies affected nodes.

Transfer a packet
5
Questions in discussion board

• Question about Fabric Manager
• How to make Fabric Manager robust ?
• How to build scalable Fabric Manager ?
• ? Redundant deployment or cluster Fabric manager
  could be solution
• Question about reality of base-line tree topology
• Is the tree topology common in real world ?
• ? Yes, multi-rooted tree topology has been a
  traditional topology in data center.

A scalable, commodity data center network
architecture Mohammad La-Fares and et el,
SIGCOMM 08
6
Discussion points

• Is the PortLand applicable to other topology ?
• ? The Idea of central ARP management could be
  applicable.? To solve forwarding loop problem,
  TRILL header-like method would be necessary.?
  The benefits of PMAC ? It would require larger
  size of forwarding table.
• Question about benefits from VM migration
• VM migration helps to reduce traffic going
  through aggregate/core switches.
• How about user requirement change?
• How about power consumption ?
• ETC
• Feasibility to mimic PortLand on layer 3
  protocol.
• How about using pseudo IP ?
• Delay to boot up whole data center