1 00:00:00,001 --> 00:00:03,723 [No Audio] 2 00:00:03,734 --> 00:00:06,934 As we know, students have limited computing 3 00:00:06,934 --> 00:00:09,901 infrastructure. If you are using Kubernetes for 4 00:00:09,901 --> 00:00:12,767 learning purpose, and you end up running 3 or 5 00:00:12,767 --> 00:00:15,634 more VMs simultaneously, your system may not 6 00:00:15,634 --> 00:00:19,401 remain in its best shape, which is not a happy day. 7 00:00:20,101 --> 00:00:23,001 On the other hand, industries have much more 8 00:00:23,001 --> 00:00:25,767 sophisticated infrastructure like servers, which 9 00:00:25,767 --> 00:00:29,067 can hold large number of VMs. And even if they run 10 00:00:29,067 --> 00:00:32,667 out of resources, they can always buy or rent more. 11 00:00:33,667 --> 00:00:37,067 So, the solution should be simple, right? 12 00:00:37,067 --> 00:00:41,167 If you are a student, use Cloud, if you are a 13 00:00:41,167 --> 00:00:43,434 professional with industrial infrastructure, 14 00:00:43,934 --> 00:00:48,067 use local installation. Well, it's not that simple. 15 00:00:48,534 --> 00:00:52,601 Because there is a thing called hosted or managed Kubernetes as well. 16 00:00:53,101 --> 00:00:54,701 It is provided by leading 17 00:00:54,701 --> 00:00:58,901 public Cloud providers like GCP, AWS, or Azure, 18 00:00:59,001 --> 00:01:01,601 and it gives all the facilities of Kubernetes 19 00:01:01,601 --> 00:01:04,400 cluster without the pain of managing it. 20 00:01:04,766 --> 00:01:09,534 So before deciding whether it is right solution for us or not, 21 00:01:09,534 --> 00:01:12,767 let's look at its pros and cons. 22 00:01:13,567 --> 00:01:16,834 It provides quick and limitless scaling very efficiently, 23 00:01:16,834 --> 00:01:19,567 as seen in Pokemon Go case study. 24 00:01:20,067 --> 00:01:23,267 But if your containers demand specific infrastructure, like 25 00:01:23,267 --> 00:01:26,667 certain GPU or something like that, it may not be 26 00:01:26,667 --> 00:01:30,601 available in your region. In which case, you may 27 00:01:30,601 --> 00:01:33,867 have to resort to private infrastructure. You need 28 00:01:33,867 --> 00:01:36,167 to hire less people to manage your cluster and 29 00:01:36,167 --> 00:01:38,834 leave all the hassle to cloud providers. 30 00:01:39,401 --> 00:01:43,867 But if the existing staff is unaware of this, or is not 31 00:01:43,867 --> 00:01:47,734 skilled enough, they may have to do a minor skill uplifting. 32 00:01:47,734 --> 00:01:50,101 Pay as you go and dynamic scaling 33 00:01:50,101 --> 00:01:52,434 reduces the risk of unwanted infrastructure 34 00:01:52,434 --> 00:01:56,234 investments. But carelessness like leaving unused 35 00:01:56,234 --> 00:01:59,201 nodes open can bring tears to wallets. 36 00:01:59,601 --> 00:02:03,501 High availability, Load balancing, and Monitoring are 37 00:02:03,501 --> 00:02:06,667 mostly one click away due to Cloud services. 38 00:02:06,834 --> 00:02:09,300 But you may have to go through an unavoidable 39 00:02:09,300 --> 00:02:12,401 migration process, you can arguably get more 40 00:02:12,401 --> 00:02:15,634 reliable performance with more security risks, 41 00:02:15,634 --> 00:02:19,267 but again, both of these aspects are debatable. 42 00:02:19,267 --> 00:02:23,267 So, for someone who doesn't want to scale their staff and 43 00:02:23,267 --> 00:02:26,101 wants quick scale up with less risks, managed 44 00:02:26,101 --> 00:02:29,034 Kubernetes might be the best choice. But for 45 00:02:29,034 --> 00:02:31,767 someone who wants to keep their data absolutely 46 00:02:31,767 --> 00:02:34,134 safe, and doesn't want to migrate private 47 00:02:34,134 --> 00:02:36,601 infrastructure would be a better option. 48 00:02:37,001 --> 00:02:40,834 But as always, for you as a student, learning managing 49 00:02:40,834 --> 00:02:44,801 Kubernetes will definitely be a great asset and a skill. 50 00:02:44,801 --> 00:02:46,734 So let's get started.