Ravindra BagaleCourses & study guides मराठी Track your progress

DynamoDB DAX Caching and Read Acceleration

By , trainer LinkedIn Last updated:

DAX is a sports car — useless if the road (keys) is wrong. Mexico City gaming lead Carlos Rivera and Kochi media engineer Priya Nair said: We added DAX before fixing a hot PK — cache stampeded on miss. Cache hot Gets after Query shapes are right. Capacity: capacity. Modeling: modeling.

What you'll learn in this chapter

  • Explain DAX simply
  • Know consistency trade-offs
  • Decide DAX vs cheaper caches
  • Industry map
  • Lab mindset + cleanup

1. Everyday example: counter display

A shop counter shows stock from a whiteboard (cache). Board can lag; cashier still checks the notebook (DynamoDB) for money. Meera caches public doctor profiles; never caches prescription ledgers.

2. Vertical steps

Why -> What -> How -> Lab -> CleanupWhyWhatHowLabCleanup
Cleanup

3. Why do we need this?

Why do we need this?

Hot GetItem paths can dominate RCU and latency. After keys are correct, a cache can absorb repeated reads. DAX is DynamoDB-aware cache — but node hours cost even idle.

4. What is it?

What is it?

DAX: managed in-memory cluster in front of DynamoDB. Item and query caches; consistency nuances vs direct DynamoDB strong reads. Alternatives: API Gateway/CloudFront, ElastiCache, or app caches.

5. How to do it (step by step)

How to do it (step by step)

  1. Prove the hot path is Get/Query — not Scan.
  2. Fix keys/hot partitions first.
  3. Estimate QPS and item size for cluster sizing talk.
  4. Decide: DAX vs CDN/API cache vs ElastiCache.
  5. Plan miss path and invalidation/TTL mental model.
  6. Never put money-critical strong-read ledgers on stale cache alone.

6. Commands with output and meaning

# DAX is managed separately from the table API.
aws dax describe-clusters --region ap-south-1
# Output: Clusters list (empty if none). Meaning: optional infra —
# create only after a written hot-Get justification; delete lab clusters same day.

describe-clusters shows spend. Create only with justification; lab clusters deleted same day.

7. Industry stories

  • Gaming (Carlos): avatar/profile Gets via DAX; wallet direct.
  • Media (Priya): public metadata CDN; personalized via DynamoDB.
  • Retail (Aisha): product card cache; checkout strong path.
  • SaaS (Omar): feature-flag Gets cached; billing direct.
  • IoT (Mei): prefer rollups/S3 over DAX for time-series.
  • Bank (Lucas): no DAX on ledger — strong correctness.

8. Lab

How to do it (step by step)

  1. Write: three reads OK to cache; two must be strong.
  2. Optional: describe-clusters; if you create a tiny lab cluster, delete same day.
  3. Re-measure without cache after key fix — often enough.

Cleanup (stop billing)

  1. Delete lab resources you created.
  2. Confirm with describe that names are gone.

9. Ops checklist before you ship

  1. Confirm TTL attribute name and units (epoch seconds).
  2. Decide soft-delete vs hard-delete vs archive-to-S3.
  3. Enable PITR on any table you cannot rebuild from source.
  4. Document RPO/RTO and who runs restore drills.
  5. If multi-region: pick write stickiness and conflict rules.
  6. If DAX: prove cache-miss path and warm-up plan.
  7. Tag Owner + Expiry on every lab resource.
  8. Budget alert in the account before load tests.

10. Copy-paste DynamoDB CLI starters (batch 2)

export AWS_DEFAULT_REGION=ap-south-1

# TTL status
aws dynamodb describe-time-to-live --table-name ShopApp

# PITR / continuous backups
aws dynamodb describe-continuous-backups --table-name ShopApp

# On-demand backup
aws dynamodb create-backup --table-name ShopApp --backup-name lab-shop-$(date +%Y%m%d)

# List backups
aws dynamodb list-backups --table-name ShopApp

# Global table description (v2)
aws dynamodb describe-table --table-name ShopApp --query 'Table.Replicas'

11. More industry stories

  • Retail (Aisha Khan, Dubai): TTL clears abandoned carts; soft-delete keeps 30-day undo.
  • Bank (Lucas Meyer, Berlin): PITR on ledger table; restore drill every quarter.
  • Travel (Omar Haddad, Amman): Global Tables for catalog reads; inventory writes sticky.
  • Gaming (Carlos Rivera, Mexico City): DAX for hot profile Gets; never cache money ledgers.
  • Social (Priya Nair, Kochi): adjacency edges for follows; sparse GSI for unread inbox.
  • IoT (Mei Lin, Shenzhen): device#day PK + TTL 14 days; rollups to S3.

12. Glossary (batch 2)

TermMeaning
TTLTime to Live — attribute with epoch seconds; DynamoDB deletes later
Soft deleteFlag deleted=true; item remains until TTL/archive
PITRPoint-in-time recovery — continuous backups window
Global TableMulti-region replicated DynamoDB table
LWWLast-writer-wins conflict resolution on replicas
DAXDynamoDB Accelerator — in-memory cache cluster
Adjacency listEdge items modeling relationships in one table
Fan-outWrite once; push copies to many feeds/inboxes

13. Practice: explain out loud

  1. 60s: TTL vs soft-delete vs S3 archive.
  2. 60s: PITR vs on-demand backup.
  3. 60s: when Global Tables help and hurt.
  4. 60s: DAX consistency trade-off.
  5. 60s: CartLite checkout keys in one breath.
  6. 60s: follow edge vs feed item.
  7. Record once; cut filler; record again.

14. Safe lab hygiene

  • Use ap-south-1 labs with a budget alert.
  • Prefix lab tables with lab- and your initials.
  • Delete replicas, DAX clusters, and backups you created for labs.
  • Never enable Global Tables on production without a write plan.
  • Tag every resource Owner + Expiry.
  • Export/import labs can create large S3 bills — delete prefixes.

16. Delete / archive decision pad

NeedPreferWhy
Undo 7-30 daysSoft-delete flagUser restore
Hard expiry, no readTTL attributeAuto delete later
Compliance keep coldExport/archive to S3Cheap long store
Mistake recoveryPITR / backupTable-level restore

17. Ten oral drill questions

  1. What units does DynamoDB TTL use?
  2. Does TTL delete instantly?
  3. PITR window typical length?
  4. Global Tables conflict default?
  5. When sticky writes matter?
  6. DAX vs API Gateway cache?
  7. Cart PK for a shopper?
  8. Order lines under which PK?
  9. Follow edge SK example?
  10. Unread inbox sparse GSI idea?

18. RPO / RTO mini workshop

  1. Write the worst data loss you can accept (RPO) in minutes.
  2. Write how long restore may take (RTO) before business pain.
  3. Map: on-demand backup vs PITR vs multi-region replica.
  4. Schedule a restore drill on a lab table this month.
  5. Name the person who runs the drill.

19. Cost watch list (batch 2)

  • PITR storage and on-demand backup storage.
  • Global Tables: replicated WCUs in every region.
  • DAX node hours even at idle.
  • Export to S3 + S3 storage + Athena later.
  • Streams + Lambda still bill during TTL deletes.

20. Project Definition of Done

  1. Access patterns written and peer-reviewed.
  2. Sample items drawn for happy path + one failure.
  3. Lab table created, exercised, deleted (or tagged Expiry).
  4. At least one conditional write or transaction exercised.
  5. Cleanup commands pasted in your notes.
  6. One interview answer recorded out loud.
  • TTL hygiene → Backup/PITR for “oops” recovery.
  • Global Tables → sticky writes from capacity/hot lessons.
  • DAX → only after Query shapes are correct.
  • P1 CartLite reuses modeling + transactions.
  • P2 Social reuses adjacency + Streams ideas.

22. Quick self-check (yes/no)

  1. I can explain TTL delay vs instant DeleteItem.
  2. I know when soft-delete beats hard delete.
  3. I can start PITR and say what restore costs in time.
  4. I can name one reason NOT to use Global Tables.
  5. I know DAX does not fix a bad partition key.
  6. I can sketch CartLite USER# and ORD# keys.
  7. I can sketch a follow edge item.
  8. I delete lab replicas/DAX/backups the same day.

23. Common errors and fixes

Each row is a real mistake students and juniors hit. Read the fix before you panic.

Error / symptomFix
DAX before key fixFix keys first.
Cache money ledgerStrong path.
Forget idle node costDelete labs.
Assume strong like GetItemKnow DAX consistency.
Cache ScansDo not — fix model.
No miss planDocument miss.
Stampede on expirySoft TTL / lock.
One huge node foreverRight-size / HA.
Ignore VPC/SGNetwork path.

24. Troubleshooting checklist

  1. Empty clusters: none created — OK.
  2. High latency still: check keys/hot.
  3. Stale reads: expected cache lag — design.
  4. Bill surprise: node hours.
  5. App errors: endpoint/SG/IAM.

25. FAQ

Q. Should I always use a single table?

Often yes for related access patterns; not a religion. Separate tables are fine for unrelated domains or different capacity/security needs.

Q. Do I still need RDS?

Yes for heavy joins, multi-row SQL reports, and strong relational constraints. See the intro DynamoDB vs RDS chapter.

Q. Is Scan evil?

Scan is a tool for rare admin jobs and small tables. Design so hot paths use Query/GetItem.

Q. Where are Put/Get command details?

In the CRUD survey, CLI and SDK chapters — this deep set links them instead of copying.

Q. DAX mandatory?

No — often CDN/app cache enough.

Q. Fixes hot PK?

No.

Q. Strong read?

Prefer direct DynamoDB when required.

Q. Link capacity?

Yes for cost.

26. Interview Q&A

Q. (B) DAX?

DynamoDB Accelerator in-memory cache cluster.

Q. (B) First step?

Fix access patterns/keys.

Q. (M) When not DAX?

Strong money paths; tiny traffic; Scan-heavy.

Q. (M) Idle cost?

Nodes bill even with low QPS.

Q. (S/MNC) Stampede?

Many misses after expiry hit DB together.

Q. (MNC) Alternative?

CloudFront/API cache/ElastiCache/app.

27. Tips in English, Marathi and Hindi

Ravindra Bagale's Tip (EN)

Make the Query cheap first; then decide if a cache earns its rent.

Ravindra Bagale's Tip (MR)

आधी Query स्वस्त; मग cache.

Tip (HI)

पहले Query सस्ता, फिर cache।

Try at home

  1. List 3 cacheable vs 2 non-cacheable reads.
  2. Price a small DAX node roughly (docs).
  3. Re-read modeling one page.

In short

  1. DAX after correct keys.
  2. Consistency and idle cost matter.
  3. Alternatives often enough.
  4. Delete lab clusters.

Revision, interview and comparison cards

Revise this lesson in a few minutes. Save the cards on your phone or download them as PNG files. The QR code on each card opens this lesson.

Revision card: DAX after keys. Cache on a hot PK stampsedes on miss. Fix Query shapes first; then consider DAX for hot Gets. Prove Get/Query → estimate QPS → pick tool → plan miss. Carlos Rivera caches avatars; wallet stays direct.
Revision card: DAX after keys · Download PNG · Open card page
Revision card: Cache rent. Idle nodes still bill; alternatives may be enough. DAX vs CDN/API/ElastiCache — pick for the read shape. Justify cluster → tag Owner/Expiry → delete labs same day. Priya Nair prefers CDN for public metadata.
Revision card: Cache rent · Download PNG · Open card page
Comparison: DAX vs Direct DynamoDB. DAX: Microsecond item/query cache idle node cost. Direct DynamoDB: Strong path and source of truth.
Comparison: DAX vs Direct DynamoDB · Download PNG · Open card page
Comparison: DAX vs CDN or API cache. DAX: DynamoDB-aware cache in the VPC. CDN or API cache: Public or edge HTTP cache.
Comparison: DAX vs CDN or API cache · Download PNG · Open card page
Interview card: Prove Get/Query. Caching a Scan is wasted money. Confirm hot path is GetItem/Query before DAX. Trace top endpoints → reject Scans → then cache talk.
Interview card: Prove Get/Query · Download PNG · Open card page
Interview card: Miss path. Unknown miss behaviour causes stampedes. Document TTL, stampede control, and DB fallback. Write miss doc → load test → tune.
Interview card: Miss path · Download PNG · Open card page
Interview card: Pick a cache. DAX is not the only cache. Public → CDN; personalized hot Get → DAX/ElastiCache/app. Classify reads → pick tool → cost check.
Interview card: Pick a cache · Download PNG · Open card page
Interview card: Delete lab DAX. Forgotten clusters burn money overnight. describe-clusters; delete labs; tag Expiry. Create only with ticket → same-day delete → verify empty.
Interview card: Delete lab DAX · Download PNG · Open card page

More cards: Ravi Sir's Cards.

Practice Tasks

About this scenario

Amravati Agro Market is a agriculture business in Amravati. Anjali Thakre looks after its IT and cloud work. They count money in INR and run their AWS resources in the ap-south-1 Region. In the tasks below you do real jobs for them using what this lesson teaches. All names, account IDs, IP addresses and numbers are made up for practice.

0 / 16

Easy DAX is an in-memory cache — 1

Why: Def.

What to do: Enter 1 if DAX is a DynamoDB Accelerator cache cluster.

How:
  1. Write down the numbers the task gives you.
  2. Work it out step by step on paper.
  3. Type only the final answer in the box.
  4. Click Check answer.

Easy Fix keys before DAX — 1

Why: Order.

What to do: Enter 1 if access patterns/keys come before DAX.

How:
  1. Write down the numbers the task gives you.
  2. Work it out step by step on paper.
  3. Type only the final answer in the box.
  4. Click Check answer.

Easy Idle nodes still bill — 1

Why: Cost.

What to do: Enter 1 if DAX nodes bill even when idle.

How:
  1. Write down the numbers the task gives you.
  2. Work it out step by step on paper.
  3. Type only the final answer in the box.
  4. Click Check answer.

Easy Do not cache money ledgers lightly — 1

Why: Safety.

What to do: Enter 1 if money ledgers often need direct strong reads.

How:
  1. Write down the numbers the task gives you.
  2. Work it out step by step on paper.
  3. Type only the final answer in the box.
  4. Click Check answer.

Easy Describe table (baseline)

Why: CLI.

What to do: Describe table ShopApp

How:
  1. Read the task and decide which AWS service it is about.
  2. Write aws, then the service name, then the operation (the action word).
  3. Add each needed flag (--name) followed by its value from the task.
  4. Flags can be in any order. Click Check answer.

Easy Cacheable YAML

Why: List.

What to do: YAML list cacheable with profile and flags.

How:
  1. Start from the top-level keys the task names.
  2. Indent child keys with 2 spaces (no tabs).
  3. Fill in each property with the value from the task.
  4. Click Check answer. Key order does not matter.

Easy Estimate QPS before sizing — 1

Why: Size.

What to do: Enter 1 if you estimate QPS before talking cluster size.

How:
  1. Write down the numbers the task gives you.
  2. Work it out step by step on paper.
  3. Type only the final answer in the box.
  4. Click Check answer.

Medium Stampede = many misses together — 1

Why: Risk.

What to do: Enter 1 if cache stampede is many misses hitting DB together.

How:
  1. Write down the numbers the task gives you.
  2. Work it out step by step on paper.
  3. Type only the final answer in the box.
  4. Click Check answer.

Medium DAX does not fix hot PK — 1

Why: Truth.

What to do: Enter 1 if DAX does not fix a hot partition key.

How:
  1. Write down the numbers the task gives you.
  2. Work it out step by step on paper.
  3. Type only the final answer in the box.
  4. Click Check answer.

Medium CDN/API/ElastiCache are alternatives — 1

Why: Choice.

What to do: Enter 1 if other caches can replace DAX for some reads.

How:
  1. Write down the numbers the task gives you.
  2. Work it out step by step on paper.
  3. Type only the final answer in the box.
  4. Click Check answer.

Medium Do not cache Scans as a fix — 1

Why: Anti.

What to do: Enter 1 if DAX should not be a Scan band-aid.

How:
  1. Write down the numbers the task gives you.
  2. Work it out step by step on paper.
  3. Type only the final answer in the box.
  4. Click Check answer.

Medium Do-not-cache YAML

Why: List.

What to do: YAML list nocache with ledger and checkout.

How:
  1. Start from the top-level keys the task names.
  2. Indent child keys with 2 spaces (no tabs).
  3. Fill in each property with the value from the task.
  4. Click Check answer. Key order does not matter.

Hard Plan the miss path — 1

Why: Design.

What to do: Enter 1 if you must document cache-miss behaviour.

How:
  1. Write down the numbers the task gives you.
  2. Work it out step by step on paper.
  3. Type only the final answer in the box.
  4. Click Check answer.

Hard Delete lab clusters same day — 1

Why: Hygiene.

What to do: Enter 1 if lab DAX clusters should be deleted same day.

How:
  1. Write down the numbers the task gives you.
  2. Work it out step by step on paper.
  3. Type only the final answer in the box.
  4. Click Check answer.

Hard Know consistency trade-offs — 1

Why: Model.

What to do: Enter 1 if DAX consistency can differ from strong GetItem.

How:
  1. Write down the numbers the task gives you.
  2. Work it out step by step on paper.
  3. Type only the final answer in the box.
  4. Click Check answer.

Hard DAX not mandatory — 1

Why: Choice.

What to do: Enter 1 if DAX is optional not mandatory.

How:
  1. Write down the numbers the task gives you.
  2. Work it out step by step on paper.
  3. Type only the final answer in the box.
  4. Click Check answer.

See all practice tasks in the Task Hub →