India’s Leading Domain & Web Hosting Provider
Support
Domain & Login
Blog
Google Cloud Storage, Filestore & Cloud CDN Services
Google Cloud Storage is Google Cloud's object storage service, offering Standard, Nearline, Coldline and Archive classes in regional, dual-region and multi-region locations. NTSPL designs bucket layouts, Autoclass and lifecycle policies, Filestore NFS shares and Cloud CDN or Media CDN delivery so Indian publishers, edtech and commerce teams serve content faster while paying less for storage and egress.
Google Cloud services in scope
The Problem Cloud Storage (Standard, Nearline, Coldline, Archive) and Autoclass Solve
Storage bills on Google Cloud rarely grow because of capacity alone. They grow because hot and cold objects share one Standard bucket, because every image request goes back to origin and pays internet egress, and because a single regional bucket becomes the disaster-recovery plan by default. On the delivery side, a cache-hit ratio below 70% means your origin is doing CDN work, and breaking-news or exam-result traffic then hits buckets and load balancers that were never meant to absorb it. Storage class, location type and cache key design decide both latency and cost.
NTSPL begins with Storage Insights inventory reports and Cloud CDN logs: object age, access frequency, size distribution and cache-miss reasons by URL pattern. From that we choose between Autoclass and explicit lifecycle rules per bucket, decide which data justifies the IN dual-region pairing of Mumbai and Delhi with turbo replication, and rebuild cache keys, TTLs and signed-URL policies so the edge, not the origin, serves the long tail.
When Cloud Storage, Filestore and Cloud CDN Fit Your Content Workload
Signals that point to storage & content delivery on Google Cloud
- More than 60% of objects in your Standard buckets have not been read in 90 days, according to Storage Insights or access logs.
- Internet egress from Cloud Storage or Compute Engine is your second-largest Google Cloud line item after compute.
- Origin CPU or bucket request rates spike during breaking news, results days or sales, while cache-hit ratio stays under 70%.
- Your DR plan for content is a nightly gsutil copy to another region with no tested recovery point.
- A legacy CMS or media workflow needs a shared NFS mount and is still running on a self-managed NFS VM.
When a different approach is better
- Transactional data needing row-level updates and queries belongs in Cloud SQL, AlloyDB or Firestore, not as JSON objects rewritten in a bucket.
- Low-latency block storage for a single VM's database is a job for Hyperdisk, not Filestore or Cloud Storage FUSE.
- Personalized HTML that differs per logged-in user gains little from Cloud CDN; render it on Cloud Run and cache only fragments and static assets.
Storage and Delivery Building Blocks: Storage Classes, Autoclass, Dual-Region, Filestore, Cloud CDN and Media CDN
| Service | Role in the architecture | Choose it when |
|---|---|---|
| Cloud Storage (Standard, Nearline, Coldline, Archive) | Stores objects in one namespace with per-object storage classes whose minimum storage durations are 0, 30, 90 and 365 days respectively, trading cheaper capacity for retrieval fees. | Access frequency varies sharply by object age, such as article images, lecture recordings or order invoices. |
| Autoclass | Moves each object between storage classes automatically based on its own access pattern, without early-deletion or retrieval charges, for a small per-object management fee. | Access patterns are unpredictable or mixed within a bucket, so fixed age-based lifecycle rules would misclassify hot objects. |
| Dual-region buckets with turbo replication | Replicates objects across two regions, including the IN pairing of Mumbai (asia-south1) and Delhi (asia-south2), with an optional turbo replication target of 15 minutes. | Content must stay within India yet survive a full regional outage with a defined recovery point. |
| Filestore | Managed NFSv3 file storage in Basic, Zonal, Regional and Enterprise tiers for applications that need POSIX semantics and shared mounts. | Legacy CMS, transcoding or rendering tools expect a mounted filesystem and cannot be rewritten to use object APIs. |
| Cloud CDN | Caches content from Cloud Storage backend buckets or load-balancer backends at Google's edge, with configurable cache modes, TTLs, negative caching and signed URLs or cookies. | Web assets, images and API responses are requested repeatedly by users spread across Indian cities. |
| Media CDN | Google's media-delivery platform for large-scale video streaming and large file downloads, with origin shielding and per-route cache policies. | Monthly video or download egress runs into hundreds of terabytes or live events create concurrent-viewer spikes. |
Dual-region origin with edge-cached delivery
Dual-region origin with edge-cached delivery
Readers resolve through Cloud DNS to a global external Application Load Balancer, where Cloud CDN serves cached objects and forwards misses to a backend bucket in the IN dual-region or to a Cloud Run rendering service. Editorial and media uploads land in the same dual-region bucket, are transcoded or resized, and Autoclass shifts ageing objects to colder classes while turbo replication keeps Delhi within 15 minutes of Mumbai.
Select any component
Dual-region origin with edge-cached delivery: click any of its 10 components to see its role and the design choice NTSPL makes for it, or trace the 6-step primary path.
Bucket and Cache Engineering Decisions: Location Type, Class Transitions and Cache Keys
Regional, IN dual-region or multi-region location
Regional buckets in asia-south1 are cheapest and fine when the bucket only feeds processing in the same region. Choose the IN dual-region when content must stay in India and survive a Mumbai outage; enable turbo replication only where a 15-minute RPO is contractual, because it adds a per-GB replication charge. Avoid the ASIA multi-region for Indian personal data, since its member regions sit outside India and complicate DPDP Act data-location commitments.
Autoclass versus explicit lifecycle rules
Autoclass fits media libraries where a five-year-old photo can suddenly trend, because it promotes objects back to Standard on access without retrieval fees. Explicit lifecycle rules are cheaper for append-only data like logs, backups and invoices, where age predicts access exactly: Standard for 30 days, Coldline at 30, Archive at 365, delete at seven years. Buckets dominated by objects under 128 KiB need care, because Autoclass treats tiny objects differently and per-object fees can outweigh savings.
Cache key, TTL and invalidation design
Most low hit ratios come from cache keys that include utm, fbclid or session parameters. We allowlist only parameters that change the response, such as width and format. Versioned asset URLs get one-year max-age so invalidation is never needed; article pages get 60-120 seconds with serve-while-stale. Cache invalidations are rate-limited, so editorial takedowns use path-pattern invalidations from a single runbook rather than thousands of individual URLs.
Cloud CDN or Media CDN
Cloud CDN integrates with the Application Load Balancer and suits web assets, images and downloads under a few hundred terabytes a month. Media CDN is built for streaming: segment-aware caching, origin shielding and route-level policies for HLS and DASH. We move video to Media CDN when monthly video egress exceeds roughly 200 TB or live events push concurrent viewers above 100,000, and keep images and pages on Cloud CDN behind the same hostnames.
Filestore tier and when to avoid NFS
Filestore Basic suits departmental shares where a short outage is tolerable. Zonal delivers higher performance for render or transcode scratch space. Regional and Enterprise tiers replicate across zones for CMS shares that cannot go down. We still push teams to move write-once media into Cloud Storage and mount it with Cloud Storage FUSE where read-mostly, because Filestore capacity is provisioned and billed whether used or not.
How NTSPL Re-Platforms a Bucket Estate and Puts Cloud CDN in Front of It
Storage Insights inventory
Enable inventory reports on every bucket and load them into BigQuery to profile object counts, sizes, ages and last-access patterns.
Bucket re-layout and location choice
Split mixed buckets by data type, create regional or IN dual-region buckets as decided, and copy objects with Storage Transfer Service while preserving metadata.
Class policy enforcement
Turn on Autoclass for media buckets and write explicit lifecycle rules for logs, backups and invoices, with soft delete windows agreed with legal.
Backend buckets and Cloud CDN enablement
Attach buckets to the global Application Load Balancer as backend buckets, enable Cloud CDN with tuned cache keys, and switch DNS in stages per hostname.
Media CDN for video routes
Configure Media CDN service routes for HLS and DASH paths with origin shielding and token authentication, then migrate player URLs.
Cache-hit tuning sprint
Review two weeks of CDN logs, fix the top 20 miss-causing URL patterns, and lock target hit ratios into Cloud Monitoring alerts.
What you receive
Securing Buckets and Edge Delivery: Public Access Prevention, Signed URLs and Bucket Lock
Storage breaches on Google Cloud are almost always configuration: an allUsers binding left on a bucket, legacy ACLs granting object-level access, or signed URLs that never expire. NTSPL enforces public access prevention at the organization level, removes fine-grained ACLs through uniform bucket-level access, issues short-lived signed URLs or cookies from Cloud CDN for paid content, and uses Bucket Lock where records must be immutable for regulatory retention.
- Organization policy storage.publicAccessPrevention so no bucket can be exposed to allUsers
- Uniform bucket-level access to remove object ACLs and make IAM the single source of access
- Cloud CDN signed URLs or signed cookies with expiry of minutes, not days, for subscriber content
- Bucket Lock retention policies for invoices or records that must be write-once for a fixed period
- VPC Service Controls perimeter around buckets holding personal data, blocking copies to outside projects
Cloud Storage and CDN Pricing Levers: Class Mix, Retrieval Fees, Cache Fill and Egress
Cloud Storage charges for capacity per GB-month by class and location, operations by class (Class A writes cost more than Class B reads), retrieval fees for Nearline, Coldline and Archive, and network egress. Cloud CDN adds cache egress per GB, cache fill from origin and HTTP request charges, but cache egress is priced below origin internet egress. Hit ratio and class mix therefore decide most of the bill.
- Raise cache-hit ratio from 60-70% to above 90% by cleaning cache keys and lengthening TTLs on versioned assets
- Autoclass on unpredictable media to avoid retrieval fees while still moving cold objects to Archive
- Lifecycle deletion of non-current versions after 30 days, which often removes 10-20% of capacity
- Serve AVIF and WebP renditions to cut bytes per image by 30-50% and shrink CDN egress
- Turbo replication only on buckets with a contractual 15-minute RPO, not on every dual-region bucket
Cloud CDN vs Media CDN vs a third-party CDN in front of Cloud Storage: which should I use?
Decide on content type, scale and integration. Cloud CDN is the default for websites and apps already behind a Google Cloud Application Load Balancer: images, scripts, downloads and cacheable API responses, configured in the same Terraform as your backends, with Cloud Armor in the same path. Media CDN is the choice when streaming video or very large downloads dominate, because segment-aware caching and origin shielding protect the origin during live events. A third-party CDN makes sense when you need multi-cloud origins under one edge, specialised edge compute you already depend on, or contractual delivery in markets where you have negotiated rates; you then pay Cloud Storage internet egress to that CDN unless an interconnect-based programme applies.
Key facts about storage & content delivery on Google Cloud
Storage & Content Delivery on Google Cloud in Practice
3 storage & content delivery scenarios from Digital news publishing, K-12 e-learning, Fashion e-commerce: the starting position, the services chosen, the architecture, the rollout and the measured outcome.
News publisher absorbs 14x breaking-news spikes with a 94% Cloud CDN hit ratio
K-12 e-learning company halves video delivery cost across 180,000 lessons
Fashion e-commerce brand cuts product-image load time by 61% on mobile
Genomics lab cuts whole-genome turnaround from 6 days to 30 hours with Batch on Spot
Animation studio bursts to 4,000 render cores for deadlines without buying hardware
Job portal parses 400,000 resumes a day on Cloud Run with no idle servers
Storage & Content Delivery on Google Cloud: Frequently Asked Questions
How is Cloud CDN priced compared with serving directly from Cloud Storage?
Cloud CDN charges cache egress per GB, cache fill per GB when the edge fetches from origin, and a per-request fee for HTTP lookups. Cache egress is priced below standard internet egress from Cloud Storage, so savings scale with hit ratio. At a 90% hit ratio, only one byte in ten is billed as origin fill, which is why NTSPL focuses first on cache key and TTL tuning.
Can we keep content in India and still have disaster recovery?
Yes. Cloud Storage offers the IN configurable dual-region, which pairs Mumbai (asia-south1) and Delhi (asia-south2), so both copies stay in India. Turbo replication adds a target of replicating newly written objects within 15 minutes. This supports DPDP Act 2023 data-location commitments to customers and gives a regional-outage recovery point without maintaining copy jobs.
What happens to minimum storage duration charges if we turn on Autoclass?
Objects managed by Autoclass are not charged early-deletion or retrieval fees when they move between classes or are read. Instead there is a monthly management fee per 1,000 eligible objects. For buckets with millions of very small objects, compare that fee with the expected class savings first; for large media objects Autoclass almost always comes out ahead.
How do we migrate from an on-premises NAS or another cloud's object store?
Storage Transfer Service handles scheduled or one-time transfers from other clouds, HTTP sources and on-premises filesystems using transfer agents, with checksums and incremental sync. For very large one-time loads with limited bandwidth, Transfer Appliance is an option. NTSPL runs an initial bulk copy, then incremental syncs until a short read-only window for cutover, and validates object counts and checksums per prefix.
Does Filestore work with GKE and Compute Engine at the same time?
Yes. Filestore exports NFS shares that Compute Engine VMs mount directly and GKE mounts through the Filestore CSI driver as persistent volumes with ReadWriteMany access. We typically use it for shared CMS media folders or render scratch space. Keep the Filestore instance in the same region and VPC as its clients, since cross-region NFS adds latency and egress.
How do we stop competitors and scrapers hotlinking our product images through the CDN?
Combine Cloud Armor rules on the load balancer with referrer and rate-based controls, use signed URLs for high-value assets like original-resolution images, and make sure the bucket itself is not publicly readable so the only path is through the CDN. Logging to BigQuery shows which user agents and referrers consume the most egress so rules can be targeted.