Internet on Campus

At BPHC, it is possible for us to live without food for a day, but
not without internet, even for an hour. From Facebook and
club/department work to accessing academic and professional resources,
our lives on campus depend heavily on the internet. It being so crucial
to everything we do, any disruption in the net service or speed is bound
to cause an uproar and rightly so. Though complaints about the net
speed have reduced greatly (particularly after the forming of the IPC
Student Nucleus), JC Spotlight intended to cover the campus internet
scenario so that all students know of all possible perspectives about
it. This was done so that the students could take an informed and mature
course of action if some problem arises in future.

Chronology

It was 2014. Modi had just been elected Prime Minister. Those were
the times when half of the present campus crowd was studying for their
entrance exams. The campus internet speeds were agreeable to everyone.
Net wasn’t as good as one would have access to at home. Yet, it was
agreeable. Those who wanted large file downloads used to do it early in
mornings. Problems in speed started in second semester of 2014-15. A few
discussions ensued in FEG and it ended there. In the first semester of
2015-16, however, the speed fell to new lows. Everyone on campus mailed
the IPC office at least once, leading to IPC, in consultation with
Student Representatives (nominated by the then Student Union),
introducing a data cap of 800 MB per student per day. The idea behind
the introduction of the cap was to ensure that a minority of users don’t
eat up most of the bandwidth. Following an uproar on FEG and Shoutboxx,
however, the cap was removed. Rather, an 800 MB per download cap was
introduced. Slow net still ensued though. In February 2016, a new
internet policy was adopted, with a 2 GB data cap per person per day.
Since the cap was within the acceptable range of most, it wasn’t
contested.

Current Scenario

The campus has a total of 490 Mbps bandwidth. Of this, 90 Mbps is
allocated to ACAD Block and the remaining is used by the rest of the
campus. As of the date of publication of this piece, no bhavan-wise
individual allocation was in place. The data cap is 2 GB per user per
day. From 5 days before the major examinations (T1, T2 and Compre), the
cap is raised to 5 GB per day.

The Truths and Myths of Slow Net:

What changed between 2014 and now?

Honestly, we couldn’t come to a conclusion as to what really changed.
The average daily consumption of the whole campus for the month of
September 2016 was 2500 GB per day. It means that not everyone exhausts
their data cap. Yet, one of the many possible ways of explaining the
difference between 2014 and now could be that, given we are growing
technically and culturally by the semester, our data needs have been
increasing at a rate faster than the rate at which bandwidth was being
added. We don’t have consumption data of 2014 to test this hypothesis
though. There could be other reasons too.

The Cost Factor

Getting extra bandwidth at our campus is quite different from getting
it at our homes. In the urban and suburban parts of the city, the
Internet Service Providers (ISPs) will have a dense network of optic
fibres. Therefore, if we take a new connection, all he will have to do
is add a small extension from the optic fibre from the road to our
bedroom. But when new bandwidth is added at our campus given all the
existing optic fibres are in full use, it costs a lot more because the
line has to be laid from campus to the nearest beam line (which is
kilometers away).

Bandwidth vs. Latency

In reality, bandwidth (though important) is not the only factor that
determines the speed of the net. Another significant factor affecting
speed is Latency. Latency is the delay in response of the website.
Here’s a crude analogy. 100 persons are to start at main gate and get
into F108 as soon as possible. Two factors determine the rate at which
they will enter F108. One, the speed at which they can run. Two, even if
everyone runs at their peak speed and reach the door at the same time,
they will have to queue up to get in. And this queue length depends on
the width of the door. Here, the time it takes to reach the door is
equivalent to latency. And the width of the door is equivalent to the
bandwidth. Therefore, only high bandwidth may not always ensure speed.
The latency (time lag) should also be low. And this latency varies from
one website to another and depends on varied hardware factors like your
laptop, the website servers, the router etc.

Then why the Data Cap?

According to Aditya Krishna (2014 batch), student representative (who
was nominated by the then SU, not to be confused with IPC Nucleus), the
2 GB figure was arbitrary. It wasn’t based on the data consumption
statistics or optimization of data consumption. Rather, a blunt attempt
to enforce a cap. The only rationale for the cap was a conviction that a
minority of users eat into the bandwidth.

Cyberoam and Load Balancers

Many a times, it has been suggested that there could be a bad
configuration of the Cyberoam. According to Aditya Krishna, when he
talked about this to the IPC faculty, “he [said] that hiring people for
the amount they are hired here, this is probably the best level of
expertise you can expect here.”

The National Knowledge Network

The NKN is a government-run multi-gigabit network which specifically
serves the educational and research institutions of the country.
BITS-Pilani, as a university, is included in this network and gets a
bandwidth of 10 Gbps. Of this, 150 Mbps is allocated to internet use and
the remaining for resources like NPTEL. Sreekar Chigurupati (2014
batch) contacted Prof. Chittaranjan Hota in regard to this in 2015.
According to Prof. Hota, because of some architectural issues, the NKN
line cannot be shared with the Hyderabad campus. And BPHC cannot request
for a separate line for itself as we are not a separate university from
BITS-Pilani.

Going the Ethernet way

A good part of our daily consumption is Facebook. And it’s primarily
used to communicate with people within the campus. Yet, the data has to
go all the way to Fb servers and come back. Something similar happens
with downloads (though not in large scale, thanks to DC). A Linux
Repository has been created on our campus’ servers (Courtesy: Siddhanth
Kumar Patel and Srimanta Barua) so that multiple people need not
download the same update from net and waste their data. An initiative
was taken in the year 2015 by Nischay Mamidi and Rohitt Vashishtha (both
2015 batch) to make FEG and Shoutboxx posts accessible through
Ethernet, so that multiple persons needn’t load these groups. However,
it didn’t click. According to the current Internet Policy, Facebook,
Gmail, YouTube, and Wikipedia are given partiality for the speed.

Conclusion

The internet speed is satisfactory now, especially after the
constitution of student representation in form of the IPC Student
Nucleus. Some volatility is observed now and then, but given the number
of users, that is expected. JC Spotlight nevertheless covered this issue
for the sake that you get a presentation of all the discussions and
perspectives on it and in case of any issue in the future, you have most
of the facts at your disposal.

Some changes are still in requirement for optimal internet usage.
These include making use of intranet instead of internet wherever
possible and statistics based data cap. And the work is hopefully in
progress. The principle problem till last semester, according to
Saivivek Peddi (IPC Nucleus), was that IPC couldn’t understand the
student requirements. Given the Student nucleus, dialogue with IPC is
easier now.

Note: Mbps in the current article means “Megabits per second”. Same goes for Gbps.

Acknowledgements

  • Saivivek Peddi (IPC Student Nucleus Member)
  • Sreekar Chigurupati
  • Aditya Krishna (SU nominated Student Representative in 2015)
  • Nischay Ram Mamidi

Reference

Article by Priyanka Perumalla, Rohitt Vashishtha, Lipi Deepakshi
Patnaik. Originally published on SWD Chronicles on February 18th, 2017.