Unit 3: Cloud storage and advanced topics
Cloud Computing notes · PTU syllabus (PGCA1937)
On this page
- Unit summary
- Provisioning cloud storage
- Virtual storage containers
- Cloud storage interoperability: CDMI and OCCI
- Database storage and resource management
- Energy efficiency in the cloud
- Market-oriented cloud computing
- Federated cloud computing
- Mobile cloud computing
- Fog computing
- Big data analytics and IoT basics
- Key terms
- Quick revision
- Important questions
Unit summary
Storage and resources are the heart of cloud services, and new models extend the cloud to markets, federations, mobiles and the edge. This unit covers provisioning cloud storage, virtual storage containers, interoperability with CDMI and OCCI, database storage and resource management, energy efficiency, market-oriented and federated clouds, mobile cloud computing, fog computing, and the basics of big data analytics and IoT.
After this unit you can
- Provision cloud storage and explain its interoperability standards
- Explain database storage, resource management and energy efficiency
- Explain market-oriented and federated cloud computing
- Explain mobile cloud, fog computing, big data and IoT
PTU syllabus topics
- Provisioning cloud storage
- virtual storage containers
- cloud storage interoperability (CDMI, OCCI)
- database storage and resource management
- energy efficiency
- market-oriented and federated cloud computing
- mobile cloud computing
- fog computing
- big data analytics and IoT basics
Topic 1
Provisioning cloud storage
Object storage
Objects with metadata in buckets, accessed over HTTP APIs; highly scalable
Amazon S3, Azure Blob, Google Cloud Storage
Block storage
Raw volumes attached to virtual machines
Amazon EBS, Azure Disks
File storage
Shared file systems over NFS or SMB
Amazon EFS, Azure Files
Archive storage
Cheap, slow-retrieval cold storage
S3 Glacier, Azure Archive
- Provisioning: choose type, region, capacity and performance tier; set access policies, encryption, versioning and lifecycle rules; storage grows automatically for object storage.
Example
Choosing storage: photos for a website go in object storage (S3); a database's disk uses block storage (EBS); shared office files use file storage (EFS or Azure Files).
Exam tip
Object, block and file storage — match each to its use.
Topic 2
Virtual storage containers
- A virtual storage container is a logical unit of storage — a bucket, container or volume — that hides the physical devices behind it. Containers carry their own access controls, quotas, metadata and lifecycle policies, and can be replicated across zones or regions.
Example
A bucket for each purpose: a company creates separate buckets for public images, private documents and backups, each with its own access policy and storage class.
Exam tip
Containers organise storage and apply policies at the container level.
Topic 3
Cloud storage interoperability: CDMI and OCCI
Full form
Cloud Data Management Interface
Open Cloud Computing Interface
Body
SNIA (Storage Networking Industry Association); also ISO/IEC 17826
Open Grid Forum
Scope
Standard RESTful interface to create, read, update and delete data in cloud storage, with metadata and containers
RESTful API to manage IaaS resources — compute, storage, network
Goal
Move data between storage providers without lock-in
Manage infrastructure across clouds with one API
- In practice, the Amazon S3 API has become a de facto standard that many storage systems implement.
Example
Why standards help: with CDMI, a backup tool can manage data in any compliant cloud storage service through one interface, avoiding lock-in to a single provider's API.
Exam tip
CDMI (SNIA) is for storage management; OCCI (OGF) is for managing compute and infrastructure.
Topic 4
Database storage and resource management
Relational (DBaaS)
Managed SQL databases with automated backups, patching and scaling
Amazon RDS, Azure SQL Database, Cloud SQL
NoSQL
Key-value, document, wide-column or graph stores scaling horizontally
DynamoDB, Cosmos DB, Bigtable
Data warehouse
Columnar analytics stores
Redshift, BigQuery, Synapse
- Resource management: scheduling and allocating VMs, storage and bandwidth among tenants — admission control, load balancing, VM placement and consolidation, live migration — to meet SLAs at minimum cost.
Example
Choosing a cloud database: an online store keeps orders in a managed relational database (Amazon RDS) and product reviews in a NoSQL database (DynamoDB) that scales easily with large volumes.
Exam tip
Relational DBaaS for structured, transactional data; NoSQL for flexible, large-scale data.
Topic 5
Energy efficiency in the cloud
- Data centres consume large amounts of electricity for servers and cooling.
PUE
Total facility energy ÷ IT equipment energy
Ideal
1.0 (all energy used by IT equipment)
Example
1,500 kWh total for 1,200 kWh IT load → PUE 1.25
- Techniques: VM consolidation and switching off idle servers, dynamic voltage and frequency scaling, efficient cooling (free air, liquid cooling), renewable energy, and carbon-aware scheduling.
Example
Measuring efficiency: a data centre with PUE = 1.2 uses 1.2 kWh in total for every 1 kWh delivered to servers — the extra 0.2 for cooling and power losses. Older data centres often have PUE around 2.
Exam tip
Power usage effectiveness (PUE) = total facility energy ÷ IT equipment energy; closer to 1 is better.
Topic 6
Market-oriented cloud computing
- Resources are traded like commodities: providers set prices and customers choose based on cost and quality of service, often through brokers and exchanges.
- Pricing models
- On-demand, reserved, spot (auction-like spare capacity), savings plans
- SLA-based admission
- Accept requests only if SLAs can be met
- Brokers
- Select the best provider for each request
- Accounting and billing
- Metering and charging
- Market exchange
- Matching buyers and sellers of resources
Example
Spot pricing: a company runs non-urgent video processing on spare cloud capacity (spot instances) at a large discount, accepting that the provider may reclaim the servers at short notice.
Exam tip
Pricing: on-demand, reserved (commitment discount) and spot (spare capacity).
Topic 7
Federated cloud computing
- A federated cloud (inter-cloud) links several providers' clouds so workloads and data can move among them — for capacity bursting, geographic reach, resilience or compliance.
- 1Local cloud reaches capacity
- 2Federation broker finds partner clouds with free resources
- 3SLA and pricing are agreed
- 4Workload is deployed or migrated
- 5Usage is metered and settled
- Challenges: interoperable APIs, identity federation, data movement costs, consistent security policies. Multi-cloud management tools and Kubernetes ease portability.
Example
Bursting to another cloud: a university's private cloud is full during exam results. A federation lets it run extra workloads on a partner's or public cloud temporarily.
Exam tip
Federation enables resource sharing, reliability and avoiding vendor lock-in.
Topic 8
Mobile cloud computing
- Mobile apps offload storage and processing to the cloud — sync, backup, streaming, AI features.
- Benefits: extends battery and storage, access anywhere, easy updates. Issues: latency, connectivity, data costs, privacy — edge computing processes data closer to users.
Example
Offloading work: a phone photo app uploads images to the cloud, where powerful servers apply AI enhancements and send the result back — saving the phone's battery and processing power.
Exam tip
Benefits: extended battery life and storage. Challenges: network latency and connectivity.
Topic 9
Fog computing
- Fog computing extends cloud services to the network edge — gateways, routers and local servers — so data from devices is processed nearby before (or instead of) going to the cloud.
Cloud
Central data centres
Massive storage and analytics
Fog
Local network nodes between devices and cloud
Low latency, local aggregation, bandwidth saving
Edge
On or next to the device
Instant response, offline operation
Example
A smart traffic system processes camera feeds at roadside fog nodes to change signals in milliseconds, sending only summary statistics to the cloud.
Topic 10
Big data analytics and IoT basics
- Five Vs of big data
- Volume, velocity, variety, veracity, value
- Cloud big data services
- Hadoop and Spark clusters (EMR, Dataproc, HDInsight), data lakes, serverless analytics
- IoT
- Sensors and devices connected to the internet producing data streams
- Cloud IoT platforms
- Device registry, messaging (MQTT), rules, analytics — AWS IoT Core, Azure IoT Hub
- Combined flow
- Devices → fog gateway → cloud ingestion → storage → analytics and dashboards
Example
IoT data in the cloud: sensors in 500 delivery trucks send location and engine data every minute to the cloud, where analytics predict breakdowns and optimise routes.
Exam tip
Five Vs of big data: volume, velocity, variety, veracity and value.
Key terms
- Object storage
- Storage of data as objects with metadata in buckets
- CDMI
- SNIA standard interface for cloud data management
- PUE
- Ratio of total facility energy to IT energy
- Federated cloud
- Interconnected clouds sharing resources
- Fog computing
- Processing at network nodes between devices and the cloud
Quick revision
- Object, block, file and archive storage; provisioning; virtual storage containers.
- CDMI (SNIA) for data; OCCI (OGF) for infrastructure; S3 API.
- DBaaS, NoSQL, warehouses; resource management.
- PUE and green techniques; market-oriented pricing; federation.
- Mobile cloud; fog vs edge vs cloud; big data Vs; cloud IoT platforms.
Important exam questions
Practice questions written to the PTU exam pattern for this unit's syllabus: short answers (Section A style) and long answers (Sections B and C style).
Short-answer questions
- Q1.Distinguish object and block storage.
- Q2.What is CDMI?
- Q3.What does OCCI standardise?
- Q4.Calculate PUE for 2,000 kWh total and 1,600 kWh IT load.
- Q5.What is a spot instance?
- Q6.Distinguish fog and edge computing.
Long-answer questions
- Q1.Explain provisioning of cloud storage and interoperability standards.
- Q2.Explain database storage, resource management and energy efficiency in the cloud.
- Q3.Explain market-oriented and federated cloud computing.
- Q4.Explain mobile cloud computing, fog computing and cloud support for big data and IoT.
Stuck on this unit?
Message SBS on WhatsApp for help with Cloud Computing, or to ask about studying M.Sc IT at Synetic.
