Efficient datatraffic

From Green software

(Difference between revisions)
Jump to: navigation, search
Current revision (14:42, 16 September 2010) (edit) (undo)
(Sources)
 
Line 35: Line 35:
== Sources ==
== Sources ==
-
This best practices was recognized as such by IT professionals, described in [[Energy efficient software (Master Thesis)|''Energy efficient software'']].
+
This best practice was recognized as such by IT professionals, described in [[Energy efficient software (Master Thesis)|''Energy efficient software'']].
 +
<br />
 +
This best practice is mentioned in:<br />
 +
Petter Larsson. 2008. Energy-Efficient Software Guidelines. [http://software.intel.com/en-us/articles/energy-efficient-software-guidelines/ White Paper for the Intel Software Solutions Group].

Current revision

Sending a smaller amount of data through a communication line is likely to leave more idle time for all transfer devices, which leads to energy savings. The additional computation needed to increase data efficiency most likely in turn consumes energy.

Contents

Advantages

For mobile devices setting up a network connection might be costly and reducing the number of connections or the size of the data might decrease cost significantly. In wired networked devices this is less relevant, because often bandwidth is cheap if not free.

Drawbacks

Using this best practice in practice to gain energy efficiency can prove difficult because finding the tipping point where the added computations outweigh the smaller amount data can be very difficult to do, especially during runtime. When the wrong choice is made, even more energy inefficiency can be introduced than the general case.

Computing the tipping point can also consume a small amount of energy.

Implementations

Different techniques to make data traffic more efficient include data compression, piggybacking, proxying and more:

Piggybacking

To reduce the number of packets being sent in a full duplex communication channel, acknowledgments can be sent along with actual data. This technique is called piggybacking.

Caching

By storing a copy of often used files in a local or shared cache, data traffic can be reduced significantly. Keeping the data in a cache consistent with the actual data brings a lot of challenges, but of-the-shelf solutions already exist.

Data compression

Delta encoding

Delta encoding is a specific and widely accepted way to reduce data redundancy by sending and storing just data differences.

Cases

Efficient datatraffic has been succesfully implemented in a Logica project temp.

Intel(r) Corporation has done measurements on energy consumption of wireless transfer of a file. In the test uploading a big file uncompressed was compared to uploading the same file compressed. Different tests were done using different compress ratios. The tests were repeated where the download was measured instead of the upload. Their results showed that a compression ratio of 3x and higher result in power savings. Low compression rates (~1.2x) can result in extra overhead and energy inefficiency. Data set with a compression ration around 2.5-3.0x show a minimal difference in power consumption.

See Also

Piggy backing: http://en.wikipedia.org/wiki/Piggybacking_%28data_transmission%29

Sources

This best practice was recognized as such by IT professionals, described in Energy efficient software.
This best practice is mentioned in:
Petter Larsson. 2008. Energy-Efficient Software Guidelines. White Paper for the Intel Software Solutions Group.

Personal tools