Technical-requirements assessment template
Provide information about data ingestion types:
Data ingestion type |
Y/N |
Description |
Frequency |
Application access |
Y |
|
|
API gateway |
Y |
|
|
Data streaming |
N |
|
|
Batch process |
N |
|
|
ETL |
N |
|
|
Data import |
N |
|
|
Time series |
N |
|
|
Provide information about data consumption types:
Data consumption type |
Y/N |
Description |
Frequency |
Application access |
|
|
|
API gateway |
|
|
|
Data export |
|
|
|
Data analytics |
|
|
|
Data aggregation |
|
|
|
Reporting |
|
|
|
Search |
|
|
|
Data streaming |
|
|
|
ETL |
|
|
|
Provide data volume estimates:
Entity name |
Estimated # of records |
Record size |
Data volume |
Game Player |
1 MM |
< 1 KB |
~ 1 GB (1 MM * 1 KB) |
Game Instance |
6 MM (100K/day * 60 days) |
< 1 KB |
~ 6 GB (6 MM * 1 KB) |
Game User Mapping |
300 MM (6 MM games * 50 players) |
< 1 KB |
~ 300 GB (300 MM * 1 KB) |
Note
The period for data retention is 60 days. After 60 days, data must be stored in Amazon S3 for analytics, by using DynamoDB Time to Live (TTL) to automatically move data out of DynamoDB to Amazon S3.
Answer these questions about time patterns:
What time frame is the application available to the user (for example, 24/7 or 9 AM to 5 PM on weekdays)?
Is there a peak in usage during the day? How many hours? What is the percentage of application usage?
Specify write throughput requirements:
Entity name |
Writes/day |
Hours/day |
Writes/second |
Game Player |
10,000 updates |
18 |
< 1 |
Game Instance |
300,000 |
18 |
< 5 |
Game User Mapping |
1,800,000,000 |
18 |
~ 27.777 |
Notes
Game Player write operations: 1 percent of users update their profiles every day, so we expect 10,000 updates for 1,000,000 users.
Game Instance write operations: 100,000 games/day. For each game we have at least 3 write operations—at creation, start, and end—so the total is 300,000 write operations.
Game User Mapping write operations: 100,000 games/day for each game with 50 players. The average game duration is 30 minutes, and the gamer position is updated every 5 seconds. We estimate an average of 360 updates per gamer, so the total is 100,000 * 50 * 360 = 1,800,000,000 write operations.
Specify read throughput requirements:
Entity name |
Reads / day |
Hours / day |
Reads/sec |
Game Player |
200,000 |
18 |
~ 3 |
Game Instance |
5,000,000 |
18 |
~ 77 |
Game User Mapping |
1,800,000,000 |
18 |
~ 27.777 |
Notes
Game Player read operations: 20 percent of users start games, so 1 MM * 0.2 = 200,000.
Game Instance read operations: 100,000 games/day. For each game we have at least 1 read operation per player, and 50 players per game, so the total is 5,000,000 read operations.
Game User Mapping read operations: 100,000 games/day for 50 players. The average game duration is 30 minutes, and the gamer position is updated every 5 seconds. We estimate an average of 360 updates per gamer, and each update requires a read operation, so the total is 100,000 * 50 * 360 = 1,800,000,000 read operations.
Specify data access latency requirements:
Operation |
99 percentiles |
Maximum latency |
Read |
30 ms |
100 ms |
Write |
10 ms |
50 ms |
Specify data availability requirements:
Requirement |
Y/N |
Metric |
Notes |
High availability |
Y |
99.9% |
|
RTO |
Y |
1 hour |
Recovery time objective |
RPO |
Y |
1 hour |
Recovery point objective |
Disaster recovery |
N |
|
|
In-Region data replication |
N |
|
|
Cross-Region data replication |
N |
3 sec latency |
Which AWS Regions? |
Specify security requirements:
Requirement |
Y/N |
Notes |
Sensitive data store |
N |
Protected health information (PHI), payment card industry (PCI) information, personally identifiable information (PII)? |
Encryption at rest |
Y |
|
Encryption in transit |
Y |
|
Client-side encryption |
N |
|
Any proprietary or third- vendor encryption library |
N |
|
Data access logging |
N |
|
Data access auditing |
N |
|
