this is old, i mean rightfully so as it was created 4 yrs ago but Google has updated the whole case study everyone :) i wouldn't rely on this but very good details
Actually, according to the Solution Concept, "Mountkirk Games is building a new multiplayer game that they expect to be very popular. They plan to deploy the game’s backend on Google Kubernetes Engine so they can scale rapidly and use Google’s global load balancer" therefore, I don't think MIG's are the right solution.
@CloudAdvocate - I like your analysis and break down. I wonder if you want to redo this video as MountKirk games case study has changed. This is true for some of other case studies that you created too. Just a heads up!
7:44 I believe you got it wrong here. They mean that your data doesn't get lost while processing (for example an instance getting overwhelmed by requests), nothing to do with storage, so in this case you would send to pubsub.
11:43 also not sure about this. Pub sub retains messages for up to a week, I think when we define "slow arriving data" we wouldn't expect it to arrive a week later. Storing directly to a bucket may slow processing down.
I think you got the final answer wrong. The business case mentions explicitly that they opt for a managed NoSQL environment. Given that Spanner is a MySQL database, I do think the answer should be C. What's your take on this? Got the exam tomorrow, but this is probably the reason why I hate multiple choice questions.
Nice video. But you state to use Datastore as the DB, which makes sense for the game state use case, but time-series is specifically mentioned, which should be BigTable. So, I'd say BigTable overrides Datastore in many aspects and should be the main DB of choice?
Actually depends on the type of the data and requirement. If in the exam the question is around the huge amount of data with less latency then BigTable makes sense. If you are just looking for NoSQL then Datastore should be sufficient.
One doubt GK - During Problem discussions you said HTTPS LB and during Solution discussion HTTP LB, Could you please elaborate which LB to use. (To increase global footprints)
@@CloudAdvocate Hey GK, I want to say a big thank you as I have become a fan of your videos today. I am new to Google cloud and your sessions are the coolest yet informative.
When they say that they want to get rid of mysql is that mean that they need to get rid of sql database ? because cloud spanner is sql . Or they just want to get rid of it because it not scaling well ? Additionnally, sql database are not good for transactional processing, it is more for olap, or this rule juste apply for bigquery and bigtable ? When we speak about time series data, that mean streaming data ?
new case studies applicable from 1st may. so these case studies are now deprecated - can you make video on new case studies - Helicopter racing, EHR, Mountkirk game and terramearth
Pause the video and think how you can solve it and compare :).
Please solve few more questions on case studies it would be great help to others
1. Need to use Datastore or Firestore to store userprofile/game state
2. GCE with Shielded VMs for hardened Linux Distro
this is old, i mean rightfully so as it was created 4 yrs ago but Google has updated the whole case study everyone :) i wouldn't rely on this but very good details
Will upload new one soon.
@@CloudAdvocate Is the new updated series coming soon? Looking forward to it
Actually, according to the Solution Concept, "Mountkirk Games is building a new multiplayer game that they expect to be very
popular. They plan to deploy the game’s backend on Google Kubernetes Engine so they
can scale rapidly and use Google’s global load balancer" therefore, I don't think MIG's are the right solution.
@CloudAdvocate -
I like your analysis and break down. I wonder if you want to redo this video as MountKirk games case study has changed. This is true for some of other case studies that you created too. Just a heads up!
ruclips.net/video/sTWIlsfcUME/видео.html
7:44 I believe you got it wrong here. They mean that your data doesn't get lost while processing (for example an instance getting overwhelmed by requests), nothing to do with storage, so in this case you would send to pubsub.
Thank you!
11:43 also not sure about this. Pub sub retains messages for up to a week, I think when we define "slow arriving data" we wouldn't expect it to arrive a week later. Storing directly to a bucket may slow processing down.
It better to store it in gcs and process it later than storing it on queue..coz of cost and also you don't have to worry about 1 week retention.
Hey GK, Since the Cloud Datastore is been upgraded with firestore. Can we replace it here for the same. Btw great content :)
I think you got the final answer wrong. The business case mentions explicitly that they opt for a managed NoSQL environment. Given that Spanner is a MySQL database, I do think the answer should be C. What's your take on this? Got the exam tomorrow, but this is probably the reason why I hate multiple choice questions.
For me answer C is better then D
Nice video. But you state to use Datastore as the DB, which makes sense for the game state use case, but time-series is specifically mentioned, which should be BigTable. So, I'd say BigTable overrides Datastore in many aspects and should be the main DB of choice?
Actually depends on the type of the data and requirement. If in the exam the question is around the huge amount of data with less latency then BigTable makes sense. If you are just looking for NoSQL then Datastore should be sufficient.
Game state and user profile must be stored in a transactional database (kind of makes sense) and BigTable is not transactional.
Please do a case study with AWS... Would be of much help...
Please try to solve few more questions on case studies available openly
One doubt GK - During Problem discussions you said HTTPS LB and during Solution discussion HTTP LB, Could you please elaborate which LB to use. (To increase global footprints)
I am sorry for the confusion and good catch its External HTTP(S) LB.
Why would we put the "late data" into GCS instead of following the standard Pub/Sub -> DataFlow path?
great analysis GK ! Thank u very much.
For hardened Linux , we should use Shielded VMs
Great job! Thanks very much for the video!
Glad you liked it!
The sample question used was very simple and generic one, can you pick up some more specific questions which are more difficult
Please upload TerramEarth case study
Sure 👍 I will.
6:38 --> Do you really need a Cloud SQL, it says transactional not transactional + relational. So I think datastore would be only appropriate.
Agreed
@@CloudAdvocate Hey GK, I want to say a big thank you as I have become a fan of your videos today. I am new to Google cloud and your sessions are the coolest yet informative.
Thank you...for taking time and giving me the feedback. I appreciate it a lot.
Could you update this with updated mountkirk usecase
ruclips.net/video/sTWIlsfcUME/видео.html
When they say that they want to get rid of mysql is that mean that they need to get rid of sql database ? because cloud spanner is sql . Or they just want to get rid of it because it not scaling well ?
Additionnally, sql database are not good for transactional processing, it is more for olap, or this rule juste apply for bigquery and bigtable ?
When we speak about time series data, that mean streaming data ?
Yes for their workload its better to use NoSQL than SQL based database for better scalability.
Bigquery is good for OLAP.
new case studies applicable from 1st may. so these case studies are now deprecated - can you make video on new case studies - Helicopter racing, EHR, Mountkirk game and terramearth
Uploading Helicopter racing today :)
will the exam questions be limited to the case studies on google site or it can be new case studies in the exam?
Hi, thanks for the video. What about TerramEarth? Tks
Will work on that..thank you.
Hi .. do you provide Azure training?
Nope
Cool