1. Changes in apps being used:
   https://www.nytimes.com/interactive/2020/04/07/technology/coronavirus-internet-use.html
   Won't discuss this too much; instead will focus on the infrastructure itself.
    --> Use of online chat apps
    --> Games
    --> Netflix
    --> Duo, etc.

2. The Internet and cloud computing is incredibly elastic:
    (https://www.fastly.com/blog/how-covid-19-is-affecting-internet-performance)
    --> Elasticity both on the supply and the demand side.
    --> Supply side: been provisioned for a rainy day (slashdot effect or flashcrowd) or for the worst case during the day/week.
        --> Cloud computing has also helped with technology such as auto scaling.
        --> As a result, there's more capacity than required at any time.
    --> Demand side:
        Also, things like adaptive bitrate algorithms, degraded quality images, compressed images, progressive loading,
        degrading video quality, e.g., YouTube reducing video quality in Europe.
        (https://www.bbc.com/news/technology-51968302)
        (https://www.bloomberg.com/news/articles/2020-03-19/netflix-to-cut-streaming-traffic-in-europe-to-relieve-networks)
    --> Aside: in many scenarios in the US where network capacity is plentiful,
        it's computation rather than networking that's limiting video performance.
    --> The Internet is inherently adaptive unlike the phone lines.
    --> However:
        Spikes still happen from time to time (e.g., Amazon prime day outage in 2018), but in general there's considerably
        more excess capacity available.
    --> In other words, the Internet is becoming a fairly reliable piece of infrastructure (like electricity or the water supply)
        rather than an esoteric research curiosity.

3. Many innovations over the years have helped the Internet scale
    (https://www.fastly.com/blog/four-key-innovations-that-prepared-the-internet-for-covid-19)
    --> Better congestion control
    --> CDNs
    --> Cloud computing
    --> Better codecs, image compression, etc.
    --> Better transport 
    --> Effects seem to be regional due to the distributed nature of the Internet.
 
4. COVID causing a flattening of the network utilization where the peaks and troughs aren't that different.
    --> Also unlike physical infrastructure like cars, more use doesn't cause wear and tear.
        (although some optic fibre cables might show signs of age.)
    --> At least not compared to the time scales at which this infrastructure is replaced (every few years).
        (https://www.vox.com/recode/2020/3/25/21188391/internet-surge-traffic-coronavirus-pandemic)

5. Latency vs. throughput
    --> Latency is still an issue with video conferencing etc. Packets being stuck in a queue.
    --> A lot of other applications are throughput sensitive: Netflix, even web page loads.
        https://www.vox.com/recode/2020/3/25/21188391/internet-surge-traffic-coronavirus-pandemic

6. Fixing video conferencing remains key
    --> Because this is one case where the *content* can not simply be brought closer to the user.
    --> Unlike how content can be brought closer to the user
        in the case of video streaming, web browsing, etc.
    --> But, even here, dedicated fiber infrastructure has been helpful with video conferencing:
        private leased lines between company datacenters, relays in the cloud
        (e.g., Zoom plus recent spill over to Oracle).
    https://techcrunch.com/2020/04/28/in-surprise-choice-zoom-hitches-wagon-to-oracle-for-growing-infrastructure-needs/

7. Other non-technical problems with COVID-19 on the Internet:
    --> Information overload (infodemic) with false information etc.
        (https://www.hcde.washington.edu/news/surviving-the-coronavirus-infodemic-discussion-with-hcde-associate-professor-kate-starbird)
    --> Phishing attempts that target people's fears about covid.
        https://blogs.akamai.com/2020/04/covid-19-phishing-exploiting-a-global-pandemic.html

8. Digital divide (https://www.fastly.com/blog/digital-divide)
    --> COVID still impacting low income communities.
    --> Raising the floor of Internet bandwidth (as comcast recently did) has been helpful.
         "ISPs can, and have, made a difference."
         Similarly, Cox communications (https://newsroom.cox.com/cox_internet_changes_to_assist_students_-_remote_workers)
    --> Internet performance seems to be correlated with income levels.
    --> And some regions (e.g., Italy, UK) seem to be faring worse than others (e.g., Japan) Internet wise.
        (https://www.fastly.com/blog/how-covid-19-is-affecting-internet-performance)

9. Going forward: Everything moving online
    --> The rise of cloud computing
    --> The decline of physical computing and physical stores.
    --> More concentration of computing power in the hands of a few cloud providers.

10. Going forward: Fixing access in developing regions remains important
    --> 5G might play a role here; avoids the cost of laying down fibre.
    --> Mobile is the primary/only means of Internet access for a large chunk of the developing world,
        given that there is limited wired infrastructure.
    --> Even in the developed world, last mile WiFi access problems persist (e.g., too many people in a room)
        (https://arstechnica.com/gadgets/2020/04/remote-work-lagging-if-you-cant-plug-it-in-upgrade-to-mesh/)