Transcript presentation

Introduction to Logistical
Networking
Micah Beck, Assoc. Prof. & Director
Logistical Computing & Internetworking (LoCI) Lab
[email protected]
APAN
Advanced Networking Conf
Aug 28, 2003
Logistical Networking
Research at UTK
University of Tennessee
• Micah Beck
• James S. Plank
• Jack Dongarra
University of California,
Santa Barbara
• Rich Wolski
US Govt. Funding
• Dept. of Energy
SciDAC
• National Science
Foundation ANIR
Industry Collab.
• Yotta Yotta
Internet2
What is Logistical Networking?
• A scalable mechanism for deploying shared
storage resources throughout the network
• A general store-and-forward overlay networking
infrastructure
• A way to break transfers into segments and
employ heterogeneous network technologies on
the pieces
Why “Logistical Networking”
• Analogy to logistics in distribution of
industrial and military personnel & materiel
• Fast highways alone are not enough

Goods are also stored in warehouses for
transfer or local distribution
• Fast networks alone are not enough

Data must be stored in buffers/files for transfer
or local distribution
The Network Storage Stack
• Our adaption of the network
stack architecture for storage
• Like the IP Stack
• Each level encapsulates details
from the lower levels, while still
exposing details to higher levels
Applications
Logistical File System
Logistical Tools
L-Bone
exNode
IBP
Local Access
Physical
IBP: The Internet Backplane
Protocol
• Storage provisioned on community “depots”
• Very primitive service (similar to block service, but more
sharable)
• Goal is to be a common platform (exposed)
• Also part of end-to-end design
• Best effort service – no heroic measures
• Availability, reliability, security, performance
• Allocations are time-limited!
• Leases are respected, can be renewed
• Permanent storage is to strong to share!
Data Movers
• Module implementing standard point-tomultipoint transfer between IBP allocations
• Uniform API allows independence from the
underlying data transfer protocol
• Not every DM can apply to every transfer
• Caller responsible for determining validity
• Current options: Multi-TCP, Multi-SABUL
(reliable), UDP Multicast (unreliable)
The Network Storage Stack
LoRS: The Logistical Runtime System:
Aggregation tools and methodologies
The L-bone:
Resource Discovery
& Proximity queries
The exNode:
A data structure
for aggregation
IBP: Allocating and managing network
storage (like a network malloc)
The Logistical Backbone
(L-Bone)
• LDAP-based storage resource discovery.
• Query by capacity, network proximity,
geographical proximity, stability, etc.
• Periodic monitoring of depots.
• 20 Terabytes of shared storage. (with plans to
scale to a petabyte...)
L-Bone: August 2003
Current Storage Capacity: 20 TB
The Network Storage Stack
LoRS: The Logistical Runtime System:
Aggregation tools and methodologies
The L-bone:
Resource Discovery
& Proximity queries
The exNode:
A data structure
for aggregation
IBP: Allocating and managing network
storage (like a network malloc)
The exNode
•
•
•
•
•
•
The Network “File Descriptor
XML-based data structure/serialization
Map byte-extents to IBP buffers (or other allocations).
Allows for replication, flexible decomposition of data.
Also allows for error-correction/checksums
Arbitrary metadata.
ExNode vs inode
IBP Allocations
the network
local system
capabilities
exNode
inode
user
kernel
block addresses
disk blocks
The Network Storage Stack
LoRS: The Logistical Runtime System:
Aggregation tools and methodologies
The L-bone:
Resource Discovery
& Proximity queries
The exNode:
A data structure
for aggregation
IBP: Allocating and managing network
storage (like a network malloc)
Logistical Runtime System
Basic Primitives:
• Upload, Download, Augment, Refresh
End-to-end Services
• Checksums, Encryption, Compression
Multithreaded Transfers
Routed/Multipath
Point-to-Multipoint
Heterogeneous Multicast
Caching/Staging
Latency hiding through
aggressive prestaging
Remote database
Wide Area Network
LAN Depot
Interactive Browser
Further Advanced Capabilities
• IBP over IPv6
• Specialized DataMovers
• Aggressive UDP (SABUL)
• Added features coming soon…
• Pipelining, Authentication, RAM resources
• Disk-to-disk transfer (Fiber Channel over IP)
• Limited computation on the depot
Architecture Publications
An End-to-End Approach to Globally Scalable
Network Storage
Micah Beck, Terry Moore and James S. Plank
ACM SIGCOMM 2002 Conference, Pittsburgh, PA, USA, August 19-23
An End-to-End Approach to Globally Scalable
Programmable Networking
Micah Beck, Terry Moore and James S. Plank
Workshop on Future Directions in Network Architecture,
ACM SIGCOMM 2003, Karlsruhe, Germany, August 27
Application Publications
An Exposed Approach to Reliable Multicast in
Heterogeneous Logistical Networks
Micah Beck, Ying Ding, Erika Fuentes and Sharmila Kancherla
Workshop on Grids and Advanced Networks, Tokyo, Japan, May 12-15, 2003
Remote Visualization by Browsing Image Based Databases
with Logistical Networking
Jin Ding, Jian Huang, Micah Beck, Shaotao Liu, Terry Moore, and Stephen
Soltesz
To appear in SC 2003, Phoenix, AZ, November, 2003
Conclusions
• IBP supports a global 20 TB testbed for distributed
applications
• Transfer rates routinely exceed 100Mbps
• New Data Movers under development
• More advanced features coming soon
• Server runs on Linux/Unix/OS X platforms
• IBP Client & LoRS also on Win32, Java
http://loci.cs.utk.edu
[email protected]