I'm curious on the decision from our two engineers to handle the database on their own for many of the initial phases instead of offloading to a fully managed database service in AWS (DDB or RDS depending on needs).
What could be the reasons they decided to host them in EC2 instances? Costs? A particular DB technology not offered as a service in AWS?
The cost difference can be significant, with self-hosting on EC2 potentially saving 40-60% compared to RDS, depending on the instance type and workload. However, it requires more management effort:
Loved the story format! It keeps the article engaging :)
A question though - won't the pre-launch stack be enough for steps 1-3? I'm doing a small project myself, and we indeed chose Amplify, Lambda and RDS. What's the advantages of an EC2 instance versus Lambdas? Cost savings?
Nice presentation and explanation. I would have loved to see the full layout of the architecture at the final stage. That would help put perspective to the final solution. Nonetheless, thanks for sharing.👏🏽👏🏽👏🏽
But you lost me at the NoSQL throwaway line at the end. Since the NoSql hype rocket has come back to earth there are many ways to store and query JSON data orher than something like Mongo DB.
Add some JSON columns to MySql/Postgres, or stick it on S3.
Very insightful. Just highlighting one typo .Thereis extra word 'million' in the statement :"They used an SQL database because it was enough to handle the first 10 million users".
It should have been "They used an SQL database because it was enough to handle the first 10 users"
I am thinking of a scenario if our application is getting used globally and we our using dynamodb, how can we manage our dynamodb ? because writes can happen at any location and it should be quick and should be available at different location. can you please help me in this
I have a question about S3 and CloudFront mentioned in step 5. What could be the usecase to use both? Won't S3 be enough to host images, css and javascript?
S3 would be good enough to host content, however to reduce latency especially for cross-geographical users CDN(Content delivery network) is mostly used. CloudFront is managed CDN offering from AWS.
In many use-cases, S3 and cloudFront are used together.
Nice article!
I'm curious on the decision from our two engineers to handle the database on their own for many of the initial phases instead of offloading to a fully managed database service in AWS (DDB or RDS depending on needs).
What could be the reasons they decided to host them in EC2 instances? Costs? A particular DB technology not offered as a service in AWS?
RDS should be enough in this context. I forgot to mention it.
Thanks, Fran.
To be fair, RDS is more expensive than self-hosting DB in EC2.
How much will be the difference u think @jakub
The cost difference can be significant, with self-hosting on EC2 potentially saving 40-60% compared to RDS, depending on the instance type and workload. However, it requires more management effort:
https://www.simform.com/blog/rds-vs-ec2/
https://medium.com/amazon-web-services/by-using-postgresql-on-ec2-save-40-to-60-cost-over-aws-rds-postgresql-e96afdd2210b
Loved the story format! It keeps the article engaging :)
A question though - won't the pre-launch stack be enough for steps 1-3? I'm doing a small project myself, and we indeed chose Amplify, Lambda and RDS. What's the advantages of an EC2 instance versus Lambdas? Cost savings?
Lambdas are billed based on the execution time. So if you have infrequent and short execution time, lambdas are a good fit.
But if you have long running and CPU intensive execution, EC2 might be cheaper. Again, it depends on the use case.
It would be interesting to know more about your side project. It's pretty cool that you've something.
Thanks!
It’s the first time I’m playing with serverless, so there are interesting lessons. Up until now I only used k8s (on GCP), with microservices.
Hopefully I’ll reach the milestones you wrote about and I will gladly share my experience :)
fingers crossed :)
excellent article and well explained
thanks
Great , brilliant way of explaining.
thanks for the feedback
This was absolutely brilliant ❤️👏🏾
thanks for the feedback, Kelvin
Amazing article! Inspires me to build something and scale it out haha!
nice
Well written. Thanks for penning it down.
thanks
This is awesome! So comprehensive and engaging.
I’m struggling a bit on Step 6 though specifically the Load balancer part and diagram. Any resources I can explore to understand that in detail?
thanks, perhaps this will might be useful https://www.nginx.com/resources/glossary/load-balancing/
Nice presentation and explanation. I would have loved to see the full layout of the architecture at the final stage. That would help put perspective to the final solution. Nonetheless, thanks for sharing.👏🏽👏🏽👏🏽
A bit late to this one, very good article!
But you lost me at the NoSQL throwaway line at the end. Since the NoSql hype rocket has come back to earth there are many ways to store and query JSON data orher than something like Mongo DB.
Add some JSON columns to MySql/Postgres, or stick it on S3.
Very insightful. Just highlighting one typo .Thereis extra word 'million' in the statement :"They used an SQL database because it was enough to handle the first 10 million users".
It should have been "They used an SQL database because it was enough to handle the first 10 users"
I am thinking of a scenario if our application is getting used globally and we our using dynamodb, how can we manage our dynamodb ? because writes can happen at any location and it should be quick and should be available at different location. can you please help me in this
Nice Article! Loved the story format!
I have a question about S3 and CloudFront mentioned in step 5. What could be the usecase to use both? Won't S3 be enough to host images, css and javascript?
S3 would be good enough to host content, however to reduce latency especially for cross-geographical users CDN(Content delivery network) is mostly used. CloudFront is managed CDN offering from AWS.
In many use-cases, S3 and cloudFront are used together.
Well put. Interesting to see what problems were faced at different levels, and decision taken to overcome them. Thanks, really great article!
Great post, excellent story.