All you have got is all it takes: Doing Top-quality Computer Architecture/systems Research, Cricket, and the Indian Research Ecosystem

Before I use a letter (word), a disclaimer. This blog is written as an individual with personal opinions, and the views expressed here do not reflect the views of my employer (organization), city, state, country, and the world that I belong to.

As someone said, there are two kinds of researchers: first, who do top-quality research and contribute significantly, and second, there is no second as there is nothing called mediocre researchers. This blog is for students (undergraduate, and graduate) and anyone who aims to do top-quality research. In computer systems, and generally in Computer science, conferences are the primary venues for the dissemination of research ideas. Among conferences, there are a few flagship conferences that you will find at the CSrankings. These conferences are extremely competitive and push the state-of-the-art, and students should target these forums. Why? The program committee (a.k.a. reviewers) are the experts in their respective research themes. Most of the time, as an author you get high-quality reviews irrespective of the outcome of the paper. As a student researcher (mostly MS and Ph.D. students), publications at these venues are an indicator of top-quality research*. This is like winning the 50-over cricket World Cup or winning a tough test series abroad. It is certainly not like winning an IPL match on a flat track.

Now the question is how?

How? At a 10K feet view, It is easy to do top-quality/fantastic research and publish at the top forums if you just work hard, period! Of course, you need a research problem to solve and do all that we discussed in the previous blogs. As a recap, take a problem, and think hard (can take months to years) :)

Now the detailed answer:

1. Aim for these conferences no matter what. This is the only way you will push yourself from your comfort zone and start contributing in a meaningful way. It is hard, it will take time but it should be your priority. Once you believe then only it is possible to travel this tough road. Think about winning the tough test series abroad. So as a first step, you need “Saahas” (Hindi word).
2. Top architecture/system conference papers are 12 to 16 pages long [2][3]. Assuming you have a problem at hand, you need to build a story like a storyteller with scientific rigor. So write, present, and communicate regularly. It is a common practice to start writing a paper at the last minute and the outcome is obvious :) Most of the time it is a rejection. It does not matter how wonderful, or beautiful your idea or results are, but if the writeup is not precise then it won’t help. In terms of test match cricket, you need to play good cricket for a good part of the test match to get a winning chance on day 4 or 5. If you play well only on day 5 then it does not help. So, “you do not have to see the whole staircase, just take the first step, one step at a time”. You will see the staircase on day 5. Do not worry.
3. Easily said than done. You need to persist with grit, no matter how good or the bad situation is. Getting the motivational plots takes months, the idea takes months, and evaluation, a few more months. So in total, it takes a year or 1.5 years to get into a zone where you feel you can target the next flagship conference. Then, you or your advisor may find that the plots you generated do not make sense now, that Figure can be improved, and that plot can be improved, etc etc for k number of iterations. So, it can be an overwhelming exercise and you need to maintain your “shraddha” (Hindi word) for your work till you submit it. Most of the time, papers get rejected multiple times before they get into these forums, and the reviews may sound harsh (BTW the reviews you get are for your submitted paper and not on you :) as the review process is double-blind). So you need to keep on ducking the balls like Rahul Dravid or Pujara, even if the wickets are tumbling on the other hand; you need to wait for the loose balls and respect the good balls session after session. There is no easy way out. So “shraddha” is the key. Do not forget the ethics and integrity aspects too else you know what happens. So, you should be honest with yourself and with your work. You may get bad umpiring decisions (bad reviews) but nowadays DRS has come and things have improved a lot :) Similarly, rebuttals/revisions are quite common among flagship conferences.
4. Next, how you work daily and do small and simple things in a great way is the key. Remember it is a test match. You will get the bouncers and googlies, and once in a while, you will get a ball for your straight drive. So you have to respect every ball and be in the game every minute. Yes, you can decide your pace and strategy, and you can take your breaks too (hard breaks are key for hard work). However, when you work, you should be involved completely in your research problem to get the best insights slowly and steadily. No need to put night outs, day outs, or whatever outs like hitting a few sixes in a T20 match. Instead, keep on doing simple things. You may get tons of ideas for a given problem and the task at hand is which idea to ditch. They said you need fire in the belly to make it big, but you do not need a one-night fire to put a night out, instead, you need the “nirantar aag” (Hindi word) that burns slowly but steadily. As Harsha Bhogle often says, Australians play hard cricket and that is why they are the most difficult side to beat in the world of cricket. They put in the hard yards day in and day out (get ready for the tough day mentally too). In summary, you need your “sadhana” (Hindi word) like a sanyasi while doing research in this world of Instagram, Twitter, Facebook, and whatnot.
5. Before taking care of steps 1 to 4, you need to play first-class matches and spend hours doing net practice, which is nothing but building your skills for doing top-class research. The skills have been discussed in the previous blogs. The above four steps will come automatically if you love, yes love what you do, and you will have fun doing research and there will be no pressure, stress, etc. So it would be great if you could have intense fun with your research work. Doing systems research in small groups (two members and an advisor) also helps as you can learn from your group members and still contribute and have fun. Just take care of three “S”s.
6. Last but not least, you need to protect the above steps (one to five) from a belief system that promotes mediocre research. It is easy to get frustrated and disheartened when you see your peers having k number of papers in unranked forums and the Indian ecosystem is celebrating it. It may also happen that peers working in some other areas get papers in months and in systems it takes years. This is perfectly fine. You should not compare a T20 match with a test match. Pause, and return to your three “S”s. You may get GYAN from your peers too. This is like folks who have never reached India test team playing eleven giving GYAN on how to play cricket. As long as your advisor is with you in terms of doing top-class research then all is well. Some committee members may say, k years over, no output, no journal papers (it used to be the case 10 years before), and you need to defend by saying it is on its way. As an Indian ecosystem, it would be great to see a few flagship papers from our students and as advisors/mentors, we should value one flagship paper over five mediocre papers with no insights at all. Things have improved a lot and will improve further if you look at the venues and not the number of papers.
7. Finally, what if you fail? It is perfectly fine and the hard yards that you have put in will help you with easy acceptances at the top-tier conferences which are not flagship. In this process, you will grow with skills, maturity, and others that will help you in preparing for your future career in research. Having said that, you should not publish at any venue just to get a publication. This will nullify all the hard yards :( Make sure you take a hard break after every submission to reboot yourself, completely. A hard break usually means no internet, no research, no deadlines, and no meeting with the advisor :)

Hope this helps. It has helped me in my microscopic research career and has helped some of my mentees who made it to these flagship forums. Hope to see more rock-solid architecture/systems research from Indian students in 2024 and yes hoping the Indian cricket team will win the next World Cup and the tough test series abroad sooner rather than later. I will end with a quote from Donald Knuth, which is highly relevant in 2023-24:

“Don’t just do trendy stuff. If something is really popular, I tend to think: back off. I tell myself and my students to go with your own aesthetics, what you think is important. Don’t do what you think other people think you want to do, but what you really want to do yourself. That’s been a guiding heuristic for me all the way through.”

[1] https://csrankings.org/#/index?all&in

[2] https://www.cse.iitb.ac.in/~biswa/HPCA24.pdf

[3] https://www.cse.iitb.ac.in/~biswa/MICRO23.pdf

*Disclaimer: There are a few specific venues that are not flagship as per CSRankings but are well suited for specific research themes.