Improved data capacity using bandwidth

acceleration in HiSeasNet

Improved data capacity using bandwidth

acceleration in HiSeasNet

Steve Foley, Jon Berger, John Orcutt, Frank Vernon

Scripps Institution of Oceanography

Steve Foley, Jon Berger, John Orcutt, Frank Vernon

Scripps Institution of Oceanography

December 14, 2009AGU Fall Meeting

HiSeasNet OverviewHiSeasNet Overview Internet connectivity to University-National

Oceanographic Laboratory System (UNOLS) research vessels via satellite Low-bandwidth, high-latency, always-on 15 Intermediate-class and larger vessels 7 Full ocean for larger vessels (C-band) 8 North America coastal for smaller vessels

(Ku-band) Program began in 2002 with one ship,

added earth station, ships, and bandwidth ever since

Program operated by Scripps Institution of Oceanography (SIO) Earth station supports four satellite beams

Internet connectivity to University-National Oceanographic Laboratory System (UNOLS) research vessels via satellite Low-bandwidth, high-latency, always-on 15 Intermediate-class and larger vessels 7 Full ocean for larger vessels (C-band) 8 North America coastal for smaller vessels

(Ku-band) Program began in 2002 with one ship,

added earth station, ships, and bandwidth ever since

Program operated by Scripps Institution of Oceanography (SIO) Earth station supports four satellite beams

HiSeasNet TechnologyHiSeasNet Technology Uses marine-stabilized

antennas 2.4m dishes for larger vessels

(C-band) 1m-1.5m dishes for smaller

vessels (Ku-band) Connectivity is all IP based

64kbps to 96kbps ship-to-shore

Shared shore-to-ship links between 192kbps (3 ships) to 256kbps (5 ships)

Allows for flexibility of any type of traffic to be sent (e-mail, web, FTP, SSH, IM, VoIP, etc.)

Uses marine-stabilized antennas 2.4m dishes for larger vessels

(C-band) 1m-1.5m dishes for smaller

vessels (Ku-band) Connectivity is all IP based

64kbps to 96kbps ship-to-shore

Shared shore-to-ship links between 192kbps (3 ships) to 256kbps (5 ships)

Allows for flexibility of any type of traffic to be sent (e-mail, web, FTP, SSH, IM, VoIP, etc.)

HiSeasNet ChallengesHiSeasNet Challenges Users are demanding more services

when they are at sea Voice-over-IP Video conferencing for outreach Large data downloads of daily/hourly

satellite images Real-time collaboration with other

scientists/ships Streaming data to shore in real time

Satellite bandwidth can be expensive! Coastal ship coverage is $750/mo, global

ship coverage is $3000/mo Bulk of the traffic (~90%) is shore-to-

ship…the expensive direction

Users are demanding more services when they are at sea Voice-over-IP Video conferencing for outreach Large data downloads of daily/hourly

satellite images Real-time collaboration with other

scientists/ships Streaming data to shore in real time

Satellite bandwidth can be expensive! Coastal ship coverage is $750/mo, global

ship coverage is $3000/mo Bulk of the traffic (~90%) is shore-to-

ship…the expensive direction

How can we increase value?

How can we increase value?

When the dollars run out, what have we done to improve user experience over a fixed sized, high-delay data pipe? Use the right satellite encoding Share the expensive (shore-to-ship) route

between ships on-the-fly Quality of Service (QoS) to keep shared

link fair QoS to manage critical/real-time traffic

differently than bulk/non-real-time traffic Cache common data (Web, DNS, FTP, etc.) Compress data

Can we use network acceleration appliances?

When the dollars run out, what have we done to improve user experience over a fixed sized, high-delay data pipe? Use the right satellite encoding Share the expensive (shore-to-ship) route

between ships on-the-fly Quality of Service (QoS) to keep shared

link fair QoS to manage critical/real-time traffic

differently than bulk/non-real-time traffic Cache common data (Web, DNS, FTP, etc.) Compress data

Can we use network acceleration appliances?

Network accelerator goalsNetwork accelerator goals Increase the amount of data that can be

sent across an existing link Ex: Push 384kbps across a 256kbps link Allows for more data to be exchanged in a

given amount of time Decrease the round trip time across a link

Ex: Have a network response of 100ms instead of 550ms

Allows for quicker response times to applications/users

Faster connection setups mean better link efficiency and ultimately more data

Improve user experience QoS to make applications get the data flows

they need

Increase the amount of data that can be sent across an existing link Ex: Push 384kbps across a 256kbps link Allows for more data to be exchanged in a

given amount of time Decrease the round trip time across a link

Ex: Have a network response of 100ms instead of 550ms

Allows for quicker response times to applications/users

Faster connection setups mean better link efficiency and ultimately more data

Improve user experience QoS to make applications get the data flows

they need

Network accelerator methods

Network accelerator methods

Open TCP sessions faster by buffering more setup instead of waiting for packet exchanges – link efficiency, user experience

Adjust TCP timers to efficiently pack long-fat satellite pipe (SCPS standard) – link efficiency

TCP packet caching (some do UDP, too) – data volume

Web / DNS caching – data volume Header and payload compression on-the-fly –

data volume QoS filtering, and auto-fragmentation – user

experience, link efficiency

NOTE: Does not improve compressed or encrypted packet payloads, just their headers

Open TCP sessions faster by buffering more setup instead of waiting for packet exchanges – link efficiency, user experience

Adjust TCP timers to efficiently pack long-fat satellite pipe (SCPS standard) – link efficiency

TCP packet caching (some do UDP, too) – data volume

Web / DNS caching – data volume Header and payload compression on-the-fly –

data volume QoS filtering, and auto-fragmentation – user

experience, link efficiency

NOTE: Does not improve compressed or encrypted packet payloads, just their headers

Accelerator appliancesAccelerator appliances Expand Networks 4830 and 4930

models 4930 units run squid web proxy on-board for

smaller ships that don’t maintain one already

Supports TCP and UDP acceleration Designed for low-bandwidth, high-

latency sat links with SCPS compliance Easy to install on-path Appliance can be remotely managed

from shore Depending on data volume license,

these are about $4k per box

Expand Networks 4830 and 4930 models 4930 units run squid web proxy on-board for

smaller ships that don’t maintain one already

Supports TCP and UDP acceleration Designed for low-bandwidth, high-

latency sat links with SCPS compliance Easy to install on-path Appliance can be remotely managed

from shore Depending on data volume license,

these are about $4k per box

Measuring accelerated data

Measuring accelerated data

Accelerators have two traffic categories: “Raw” – what would be sent if there were no

accelerator installed “Accelerated” – what is actually sent on an

interface, post acceleration The difference between raw and

accelerated data is the bandwidth improvement offered by the accelerator

Accelerators have two traffic categories: “Raw” – what would be sent if there were no

accelerator installed “Accelerated” – what is actually sent on an

interface, post acceleration The difference between raw and

accelerated data is the bandwidth improvement offered by the accelerator

ResultsResults 10 months of testing done on 2 C-band

ships Atlantis

Shore-to-ship: 13% improvement Ship-to-shore: 17% improvement

Revelle Shore-to-ship: 21% improvement Ship-to-shore: 63% improvement

Bulk of total traffic (~90%) is shore-to-ship Bulk of shore-to-ship traffic is web Ship-to-shore traffic is scattered, mostly

web requests, mail, data, IM Roughly half of improvement is

compression Ships with web proxies have a 40-50% hit

rate already

10 months of testing done on 2 C-band ships Atlantis

Shore-to-ship: 13% improvement Ship-to-shore: 17% improvement

Revelle Shore-to-ship: 21% improvement Ship-to-shore: 63% improvement

Bulk of total traffic (~90%) is shore-to-ship Bulk of shore-to-ship traffic is web Ship-to-shore traffic is scattered, mostly

web requests, mail, data, IM Roughly half of improvement is

compression Ships with web proxies have a 40-50% hit

rate already

Results by applicationResults by application HTTP from ship: 290%, to ship: 20% Instant Messaging: 10-350%

Directional and based on how much video/audio gets passed along

General IM text chat seems ~35% SSH: 0-35% (varies by content and crypto) SMTP: 20-40% POP3 from ship: 0%, to ship: 85% IMAP from ship: 65%, to ship: 30% Secure POP/IMAP/HTTP: 0% (encrypted) FTP: 0% (usually compressed already) DNS from ship: 15%, to ship: 40% Streaming video from ship: 5%, to ship:

15% Streaming data (uncompressed): 130%

HTTP from ship: 290%, to ship: 20% Instant Messaging: 10-350%

Directional and based on how much video/audio gets passed along

General IM text chat seems ~35% SSH: 0-35% (varies by content and crypto) SMTP: 20-40% POP3 from ship: 0%, to ship: 85% IMAP from ship: 65%, to ship: 30% Secure POP/IMAP/HTTP: 0% (encrypted) FTP: 0% (usually compressed already) DNS from ship: 15%, to ship: 40% Streaming video from ship: 5%, to ship:

15% Streaming data (uncompressed): 130%

Results caveatsResults caveats Benefit depends heavily on traffic

patterns, therefore it can vary with: Use policies on a ship (imposed by operator) Size of ship and science party Science work being done Cruise plan (weather/heading, port time,

duration, etc.) Human behavior

Traffic conditions vary between cruises and during cruises

HiSeasNet doesn’t inspect packets deeply 10 month average includes months when

accelerator was still being tuned Recent acceleration numbers are a bit

better

Benefit depends heavily on traffic patterns, therefore it can vary with: Use policies on a ship (imposed by operator) Size of ship and science party Science work being done Cruise plan (weather/heading, port time,

duration, etc.) Human behavior

Traffic conditions vary between cruises and during cruises

HiSeasNet doesn’t inspect packets deeply 10 month average includes months when

accelerator was still being tuned Recent acceleration numbers are a bit

better

Cost analysisCost analysis Initial investment:

15% of $3000 monthly C-band satellite lease against a $4k box is ROI of 9 months

Estimated: 30% of $750 monthly Ku-band satellite lease against a $4k box is ROI of 18 months

Hidden costs: Earth station accelerator installation Maintenance per year for ~20% for hardware

and firmware support Additional license needed with more bandwidth Added hassle of one more box to

troubleshoot/fail/blame Benefits:

Retain improvement for years to come Some added QoS for better experience

Initial investment: 15% of $3000 monthly C-band satellite lease

against a $4k box is ROI of 9 months Estimated: 30% of $750 monthly Ku-band

satellite lease against a $4k box is ROI of 18 months

Hidden costs: Earth station accelerator installation Maintenance per year for ~20% for hardware

and firmware support Additional license needed with more bandwidth Added hassle of one more box to

troubleshoot/fail/blame Benefits:

Retain improvement for years to come Some added QoS for better experience

Future plansFuture plans Whole fleet is getting accelerators now

Installation is on-going over the next few months as ships go back to sea

So far data are encouraging, but most ships ships are not under normal operations right now

Smaller ships are getting accelerators with web proxies built-in

Tune accelerators better Use extra bandwidth for:

More scientific data streams to shore in real-time

More web access Smoother voice/video

Whole fleet is getting accelerators now Installation is on-going over the next few

months as ships go back to sea So far data are encouraging, but most ships

ships are not under normal operations right now

Smaller ships are getting accelerators with web proxies built-in

Tune accelerators better Use extra bandwidth for:

More scientific data streams to shore in real-time

More web access Smoother voice/video