05Design patternsHandling large blobs
05 · Reusable pattern
Handling large blobs
Use direct transfer, chunking, checksums, metadata indirection, and background processing.Lesson spine
What you need to understand.
Large-object systems split control-plane metadata from a direct, resumable byte path.
Pre-signed transfer
Create an upload session and let the client send bytes directly to object storage under narrow, expiring permission.
Multipart protocol
Upload chunks in parallel with checksums, retry parts independently, and commit a manifest only when every required part exists.
Post-processing
Publish an event after commit so scanners, transcoders, preview generators, and metadata extractors can run asynchronously.
CDN delivery
Serve immutable renditions through signed edge URLs and shield the origin from global traffic.
State synchronization
Model initiated, uploading, verifying, ready, failed, and expired so metadata never promises bytes that are not usable.
Content-defined chunking
For file synchronization and dedupe, choose boundaries from content so insertions do not shift every later chunk.
Before the boxes
Frame the decision.
What must work
Use direct transfer, chunking, checksums, metadata indirection, and background processing.
What changes the design
p50/p99 size · part count · concurrency · retention · processing fan-out
What owns the truth
Identify the component that commits authoritative state, then separate synchronous confirmation from derived work.
What stays simple
Do not add global coordination, multi-region writes, or a specialized store until a requirement earns the complexity.
Architecture map
Trace ownership, not just traffic.
Follow the decision from left to right. Every arrow should have a reason.
Obtains transfer session
Records intent
Transfers chunks
Persists bytes
Checks manifest
Commits visibility
Runs derived work
Walk one representative request across every arrow. Say whether the handoff is synchronous or asynchronous, what identity makes a retry safe, and which step changes authoritative state.
- 01
Client — Obtains transfer session Define the output contract before moving to the next owner.
- 02
Control plane — Records intent Define the output contract before moving to the next owner.
- 03
Uploader — Transfers chunks Define the output contract before moving to the next owner.
- 04
Blob store — Persists bytes Define the output contract before moving to the next owner.
- 05
Verifier — Checks manifest Define the output contract before moving to the next owner.
- 06
Metadata DB — Commits visibility Define the output contract before moving to the next owner.
- 07
Processor — Runs derived work Confirm the result and emit the evidence needed to reconcile it.
Decision table
Make the trade-offs explicit.
| Decision | Defensible position | Cost to acknowledge |
|---|---|---|
| Primary mechanism | Fixed chunks simplify dedupe; content-defined chunks preserve similarity at greater CPU cost. | The stronger guarantee usually adds coordination, latency, state, or operational work. |
| Sync vs. async | Keep only correctness-critical confirmation synchronous. Move derived views, notifications, analytics, and cleanup behind a durable boundary. | Async work needs idempotency, lag monitoring, replay, and a product definition for partial completion. |
| Simple vs. scaled | Begin with one logical owner and a clear API. Partition or replicate only the resource proven to be the first bottleneck. | Migration requires stable identities, versioned contracts, backfill, and a rollback path. |
Failure review
Design the recovery path.
Topic-specific risk
Premature visibility exposes incomplete objects; retries leave orphaned chunks.
ResponsePersist enough identity and state to distinguish retry, resume, compensation, and operator repair.
Dependency timeout
A timeout is ambiguous: the remote side may have failed, succeeded, or still be running.
ResponseUse deadlines, bounded backoff with jitter, idempotency keys, and a status or reconciliation path.
Overload or skew
Average capacity can look healthy while a tenant, key, partition, region, or expensive request saturates one owner.
ResponseExpose queue depth and hot-key share, apply backpressure, isolate tenants, and degrade optional work before correctness.
Evidence + level bar
Prove the design can be operated.
Health of the promise
Measure user-visible latency or freshness, correctness drift, saturation, retry volume, and time to recover. Alert on the failed promise—not only CPU.
Complete and clear
Finish the happy path, identify the state owner, choose reasonable building blocks, and explain one scale mechanism.
Trade-offs and failure
Separate read and write paths, define consistency, explain partitioning, and make duplicate or partial failure safe.
Evolution and operations
Discuss multi-region boundaries, migration, tenant isolation, capacity, observability, and how the architecture changes over time.
Interview language
Open the deep dive with a claim.
“For Handling large blobs, the decision I want to make explicit is this: Fixed chunks simplify dedupe; content-defined chunks preserve similarity at greater CPU cost. I’ll trace the state-changing path first, show where the result becomes durable, then test the design against the highest-risk failure and our target scale.”
08 · Retrieval check
Can you defend it without the page?
- For Handling large blobs, where is the correctness boundary and which failure would you test first?
- Which component owns committed truth, and what event or response proves the commit?
- Where is the first scaling or coordination bottleneck under the stated envelope?
- What happens after an ambiguous timeout or duplicate operation?
- Which complexity would you remove at one hundredth of the scale?