|
NetMiner 3 supports ‘Large Network Data’, let us explain more on this with four frequently-asked questions and our answers below:
- Do all licenses of NetMiner 3 support ‘large-scale
Network Data’?
No. There are five different levels of license (Tiny, Small,
Medium, Large, and Huge) with respect to the data size for
NetMiner 3 to handle. In order to handle ‘Large Network
Data’, of maximum of 100,000 nodes or ‘ Huge
Network Data’, of maximum of 1,000,000 nodes, the
appropriate level of NetMiner 3 license should be purchased.
- If I purchased the appropriate level of NetMiner 3 license
in order to meet data size of my dataset to be handled,
NetMiner 3 work on the large-scale dataset no matter what?
No. The level of license with restriction on data size
just checks the maximum nodes to be imported into NetMiner
3. However, the overall size of network data is defined
not only by the number of nodes but also by the number of
links. In order to set the upper limit of link size, NetMiner
3 defines an operational definition of ‘Sparse Network’
as the network of ‘m ≤ 10 x n’, where ‘n’
is the number of nodes and ‘m’ is the number
of links. In other words, the number of links should be
smaller than ten times the node size and the maximum size
of links is 1,000,000 for 'Large' size and 10,000,000 for
'Huge' size.
- If I purchased the appropriate level of NetMiner 3 license
in order to meet data size of dataset to be handled and
the dataset satisfies our operational ‘Large Sparse
Network’ definition, will all NetMiner algorithms
operate efficiently regardless of its time complexity?
No. NetMiner 3 guarantees for the performance of algorithms
with subquadratic time complexity where running time f(n)
is under o(n^2). Test results of subquadratic algorithms
of NetMiner 3 are listed below. Although over-quadratic
algorithms can run in NetMiner, it is not as fast and smooth
processing of ‘Large and Huge Network Data’
as compared to subquadratic algorithms.
- If I purchased the appropriate level of NetMiner 3 license
in order to meet data size of dataset to be handled and
the dataset satisfies our operational ‘Large Sparse
Network’ definition, and if I run subquadratic algorithms
of NetMiner 3, will it be guaranteed that those run with
high performance?
No. Although your system satisfies the minimum requirements
to install NetMiner 3, the system performance such as RAM
size and CPU speed inevitably affects the NetMiner 3 performance.
Therefore, although NetMiner 3 can run with lower-level
system specifications, it is recommended to have a large
size of RAM and high speed of CPU for better performance.
To sum it up, in order to get the benefit of NetMiner's handling ‘Large-scale
Network Data‘, i) users need to have appropriate level
of license and, ii) the network data is sparse (m ≤ 10
x n) and, iii) users are running subquadratic algorithms and,
iv) NetMiner 3 is installed in a computer with a large size
of RAM and high speed of CPU.
Two large-scale sample data were used for performance test.
In case of 'Large' size of network data with 100,000 nodes
and 1,000,000 links, elaspsed time and occupied memory for
overall subquadratic algorithms in NetMiner 3 were Max. 15seconds
and 1.5GB. For 'Huge' size of network data with 1,000,000
nodes and 10,000,000 links, elaspsed time and occupied memory
were 1~4 minutes and 6~10 GB according to the time complexity
of each algorithm. This test was performed in the computing
envrionment as follows:
- CPU:QuadCore Intel Xeon E5405, 2.00GHz(6*33)
- RAM:16GB (= 4GB DDR2 SDRAM 667MHz * 4)"
- OS: Microsoft Windows Server 2003 Enterprise x64 Edition
SP2
|
