Version 18 (modified by zink@cs.umass.edu, 9 years ago) (diff)

--

F. Analyze

In Section F we went through the exercise of retrieving data from iRODS to a local computer. In this Section, we will introduce two different methods that can be used to analyze the measurement data. Analysis of measurement data obtained with OMF/OML is not limited to these two methods, we simply use them for demonstration purposes.

G.1 R Scripts

One potential way to visualize the data is making use of R, which provides a visualization language. For this tutorial, we have create a set of R script, which we briefly discuss in the following.

The first script creates a plot of the RTTs for each ping that's carried out in the experiment. The following code snipped shows part of the R script that is used to plot a single ping (to 192.168.2.10).

```library(RSQLite)
con <- dbConnect(dbDriver("SQLite"), dbname = "gimi30-ping_all.sq3")
dbListTables(con)
rtt <- mydata\$rtt
pdf("gimi31_ping1.pdf")
plot(rtt,type="o",col="red",xlab="Experiment Interval",ylab="RTT (ms)")
title(main="Ping Experiment to IP address 192.168.4.10", col.main="blue", font.main=4)
```

The resulting plot is shown below.

The following script plots results from the 4th experiment we executed.

```library(RSQLite)
con <- dbConnect(dbDriver("SQLite"), dbname = "gimi20-otg-nmetrics.sq3")
dbListTables(con)
mydata <- dbGetQuery(con, "select pkt_length from otr2_udp_in where src_host='192.168.4.10'")
pkt_length <- mydata\$pkt_length/1024/1024*8
pdf("gimi31_otg1.pdf", height=5.5, width=7, pointsize=14)
plot(pkt_length,type="o",col="red",xlab="Experiment Interval",ylab="Packet size")
dev.off()
```

Both scripts are simply executed with:

```R -f <script_name>
```

The benefit of using R scripts is that they can produce graphs that can be used in documents!

The results can then be stored into iRODS using the itools presented in Section E.2.

G.2 omf_web

The second alternative we use in the tutorial is omf_web, which already has been presented in Section D.