ACID, Catching, CDN, DS

Question 1

What does ACID stands for and what does it mean?

Answer 1

A: Atomicity. Either the entire transaccion success or fails.
C: Consistency. All db rules are enforced or rollback.
I: Isolation: No transaction is affected by other in progress.
D: Durability. Once commited, it stays, even if the system crashes after.

Question 2

Cap Theorem

Answer 2

A db can only enforce two of:
Consistency
Availability
Partition-Tolerance

Question 3

What is a Data Lake, how do you scheme and query it?

Answer 3

Approach for big ammount of unstructured data. Thrown as text files (csv, json) into ex. Amazon S3.
Schemed by Amazon Glue, queried by Amazon Athena

Question 4

How does catching work?

Answer 4

Horizontally Scalled Servers
In memory (fast)
Apps with more reads than writes
Writes invalidate catching
Expiration policy
Hotspots and Cold-start is a problem

Question 5

Eviction policies for cache

Answer 5

LRU: Least Recently Used (Double Linked List)
LFU: Least Frequently used
FIFO: First in first out

Question 6

Catching Technologies

Answer 6

Memcached: In memory key/value store. Open source.
Redis: More features, snapshots, replication, transactions, pb/sub, advanced data structures.
ElasticCache AWS.

Question 7

How does CDN Work?

Answer 7

Geographically distributed edge locations.
Distributed Static Content
AWS CloudFront, Akamai

Question 8

Mention Resilency strategies

Answer 8

Replicate data as a backup
Load Balancers with Geolocation
Distributed content across AZ and regions
Over provide

Question 9

What are Distributed Storage Solutions?

Answer 9

Services for scalable, available and secure fast object storage. S3.

Question 10

How does HDFS (Haddop distributed file system) work?

Answer 10

Files are broken up into “blocks” replicated across your server. Replication is rack-aware.
A “name node” or master coordinate operations, it has backups.
Clients read from nearest replica.