Cloud Computing Assignment | March 2026
CLOUD COMPUTING
Assignment Report
─────────────────────────────
Topics Covered:
1. Cloud Benchmarking Tools (15+ Tools)
2. Cloud Marketplace & Platforms (15+ Platforms)
3. Spotify Multi-Cloud Infrastructure, Cost Model & Security
4. Cloud Migration Research – Methods, Types & IEEE Paper
5. Cloud Security Research – Algorithms, Types & Suggestions
Submitted: March 2026
Page 1 of 14
Cloud Computing Assignment | March 2026
Question 1: Cloud Benchmarking Tools
Cloud benchmarking involves measuring and comparing the performance of cloud services to evaluate
their capacity, reliability, and efficiency. Below are 15+ widely used benchmarking tools with their uses.
# Tool Name Uses & Description
1 Apache Open-source load testing tool. Tests performance of web applications,
JMeter databases, and cloud APIs. Simulates heavy load to measure response times,
throughput, and latency. Widely used for HTTP, FTP, JDBC load tests.
2 SPEC Cloud Standard Performance Evaluation Corporation's benchmark specifically
IaaS designed for cloud IaaS workloads. Measures autoscaling, VM provisioning
speed, and workload throughput. Provides vendor-neutral metrics.
3 CloudHarmo Measures and compares performance across cloud providers including AWS,
ny Azure, GCP. Tests compute, storage, network, and database speeds.
Provides historical data for trend analysis and helps in provider selection.
4 Phoronix Comprehensive open-source benchmarking platform with 200+ test profiles.
Test Suite Tests CPU, GPU, memory, disk, and network on cloud VMs. Supports
automated testing and results publishing to [Link].
5 Geekbench Cross-platform processor benchmark measuring single and multi-core CPU
performance. Used on cloud VMs to compare compute power of different
instance types. Measures integer, floating point and memory performance.
6 Sysbench Multi-threaded benchmark for MySQL/MariaDB, CPU, memory, threads and
I/O. Widely used to evaluate cloud database instance performance. Helps
tune configuration and choose right cloud DB tier.
7 FIO (Flexible Benchmarks cloud storage performance including IOPS, throughput, and
I/O Tester) latency. Tests EBS (AWS), Azure Disk, and GCP Persistent Disk with
configurable workloads. Essential for choosing the right storage class.
8 iperf3 Network bandwidth testing tool used to measure TCP/UDP throughput
between cloud instances. Tests inter-region latency, VPC network
performance, and bandwidth caps. Used in hybrid cloud connectivity testing.
9 YCSB Benchmarks cloud databases and key-value stores like DynamoDB, Cosmos
(Yahoo DB, Bigtable, Cassandra. Tests read/write/scan workloads with configurable
Cloud load patterns. Standard tool for NoSQL performance evaluation.
Serving
Benchmark)
1 Caliper Hyperledger's blockchain performance benchmarking framework. Measures
0 TPS (transactions per second), latency, and resource usage on cloud-hosted
blockchain networks like Fabric and Ethereum.
1 CloudPerf Google-developed micro-benchmarking toolkit for measuring inter-VM
1 network performance, latency, and CPU efficiency in GCP. Identifies
performance regressions across cloud configurations.
1 Benchmark Enterprise-grade database benchmarking tool supporting TPC-C, TPC-H, and
2 Factory TPC-E standards. Tests cloud RDS instances (Oracle, SQL Server, MySQL).
(Quest) Used for capacity planning and cloud migration validation.
1 k6 (Grafana) Modern open-source load testing tool using JavaScript for scripting. Tests
3 REST APIs, microservices, and cloud applications. Integrates with CI/CD
Page 2 of 14
Cloud Computing Assignment | March 2026
# Tool Name Uses & Description
pipelines and generates detailed performance metrics and trend dashboards.
1 Nench Lightweight VPS/cloud server benchmark that tests download speed, disk I/O,
4 and CPU using dd, dd seek, and wget. Quick server characterization tool
popular for evaluating low-cost cloud VM offerings.
1 CloudSpecta Independent cloud benchmarking platform that conducts standardized tests
5 tor across IaaS providers. Measures CPU, RAM, SSD, and network performance
with normalized scoring for fair provider comparisons.
1 HammerDB Open-source database load testing tool for TPC-C and TPC-H benchmarks.
6 Tests cloud databases including AWS Aurora, Azure SQL, and GCP Cloud
SQL. Helps in right-sizing database instances.
Summary of Benchmarking Categories
Category Tools
Compute / CPU Geekbench, Phoronix Test Suite, Sysbench, Nench
Network iperf3, CloudPerf, CloudSpectator
Storage / I/O FIO, Benchmark Factory
Database YCSB, Sysbench, HammerDB, Benchmark Factory
Web / API Load Apache JMeter, k6, CloudHarmony
Blockchain Caliper
Comprehensive SPEC Cloud IaaS, CloudSpectator
Page 3 of 14
Cloud Computing Assignment | March 2026
Question 2: Cloud Marketplace & Platforms
What is a Cloud Marketplace?
A cloud marketplace is a digital catalog or online store hosted by a cloud provider where customers can
discover, purchase, deploy, and manage third-party and first-party software, services, and solutions
directly within the cloud environment. It simplifies procurement by offering pre-configured software,
SaaS applications, APIs, datasets, and professional services that are ready to deploy on the underlying
cloud infrastructure.
Key Features of Cloud Marketplaces:
• One-click deployment of verified software solutions
• Consolidated billing through the cloud provider
• Pre-configured, security-vetted solutions
• Free trials and flexible pricing models (BYOL, pay-as-you-go)
• Integration with cloud-native services and IAM
# Platform Provider Key Products in Marketplace
1 AWS Amazon Web EC2, S3, RDS, Lambda, SageMaker, EKS, CloudFront,
Marketplace Services DynamoDB, Redshift, IAM
2 Microsoft Azure Microsoft Azure VMs, Blob Storage, Azure SQL, AKS, Cosmos DB,
Marketplace Azure Power BI, Azure AI, DevOps, Functions
3 Google Cloud Google Cloud Compute Engine, BigQuery, GKE, Cloud Run, Vertex AI,
Marketplace Platform Firestore, Cloud Spanner, Pub/Sub
4 IBM Cloud IBM Cloud Watson AI, IBM Kubernetes, Db2, OpenShift, Cloud
Catalog Object Storage, Event Streams, Satellite
5 Oracle Cloud Oracle Cloud Autonomous DB, OCI Compute, Oracle Analytics,
Marketplace Infrastructure Exadata, OCI Container Engine, Blockchain
6 Alibaba Cloud Alibaba Cloud ECS, OSS, ApsaraDB, MaxCompute, Function Compute,
Marketplace ARMS, Lindorm, DataWorks
7 Salesforce Salesforce CRM, Einstein AI, MuleSoft, Tableau, Marketing Cloud,
AppExchange Platform Service Cloud, CPQ
8 Red Hat Red Hat / IBM OpenShift, Ansible, RHEL, AMQ Streams, CodeReady,
Marketplace Quay, Service Mesh, Fuse
9 VMware VMware by vSphere, NSX, vSAN, Tanzu, Horizon, Carbon Black,
Marketplace Broadcom Aria Suite
1 DigitalOcean DigitalOcean Droplets, Kubernetes, Managed Databases, App
0 Marketplace Platform, Spaces, CDN, 1-Click Apps
1 Huawei Cloud Huawei Cloud ECS, OBS, DWS, CCE, ModelArts, IoT Platform,
1 Marketplace CloudTable, GaussDB
1 Tencent Cloud Tencent CVM, COS, TencentDB, TKE, SCF, CDN, Tencent AI,
2 Marketplace Cloud GAAP
1 SAP Store SAP Cloud SAP HANA, S/4HANA, BTP, Ariba, SuccessFactors,
Page 4 of 14
Cloud Computing Assignment | March 2026
# Platform Provider Key Products in Marketplace
3 Platform Concur, IBP, Analytics Cloud
1 Snowflake Snowflake Data Sharing, Streamlit Apps, Native Apps, Data Clean
4 Marketplace Data Cloud Rooms, Cortex AI
1 HashiCorp Cloud HashiCorp Terraform Cloud, Vault, Consul, Nomad, Boundary,
5 Platform Waypoint, Packer
1 Cloudflare Cloudflare Workers, R2 Storage, KV, D1 Database, Pages, Zero
6 Workers Trust, Stream, AI Gateway
Marketplace
Page 5 of 14
Cloud Computing Assignment | March 2026
Question 3: Spotify – Multi-Cloud Infrastructure, Cost Model,
Scalability & Security
Overview
Spotify is the world's largest audio streaming platform with over 600 million users and 100 million+
tracks. It migrated entirely to Google Cloud Platform (GCP) in 2016 but also leverages a multi-cloud
strategy with selective use of other providers to ensure resilience and best-of-breed services.
1. Multi-Cloud Infrastructure
Primary Cloud – Google Cloud Platform (GCP)
• Compute: Google Kubernetes Engine (GKE) manages 10,000+ microservices
• Storage: Google Cloud Storage (GCS) for audio files, metadata, and user data
• Data Warehouse: BigQuery for analytics (250TB+ of event data daily)
• ML/AI: Vertex AI and TensorFlow for personalized recommendation engine (Discover Weekly,
Daily Mix)
• Pub/Sub: Google Pub/Sub for real-time event streaming
• CDN: Deployed globally across 35+ GCP regions for low-latency audio delivery
Secondary Cloud – Multi-Cloud Elements
• Content Delivery: Akamai CDN and Fastly for edge caching of popular audio files
• Podcast Hosting: Anchor (Spotify-owned) uses AWS for some podcast transcoding pipelines
• Advertising Platform: Spotify Audience Network partially uses Azure Cognitive Services
• Payments: Stripe and Braintree APIs hosted across multiple cloud regions
Microservices Architecture
Spotify follows a cell-based microservices architecture. Each squad owns their microservice. Services
communicate via gRPC and HTTP/2. Service mesh (Envoy/Istio) manages internal traffic. Kubernetes
autoscaling ensures dynamic resource allocation based on demand.
2. Cost Model – Including Freemium
Freemium Revenue Model
Tier Features Price Revenue Source
Free Tier Ad-supported music, shuffle ₹0 Ad Revenue (~15% of total)
play on mobile, limited skips,
30-day offline
Premium Ad-free, unlimited skips, offline ₹119/ Subscription (~85% of total)
Individual mode, high-quality 320kbps month
audio, exclusive content
Premium Duo Two accounts, one household, ₹153/ Subscription
Page 6 of 14
Cloud Computing Assignment | March 2026
Tier Features Price Revenue Source
all Premium features month
Premium Up to 6 accounts, parental ₹179/ Subscription
Family controls, all Premium features month
Premium Discounted Premium for ₹59/month Subscription
Student verified students
Cloud Cost Optimization Strategies
• Committed Use Discounts (CUDs): Spotify commits to 1–3 year GCP contracts for ~30–40%
savings on compute
• Preemptible VMs: Used for batch ML training jobs, saving up to 80% on compute costs
• Autoscaling: GKE Horizontal Pod Autoscaler adjusts pod count based on real-time traffic,
avoiding over-provisioning
• Data lifecycle policies: Cold audio files moved to Nearline/Coldline storage, reducing storage
costs
• Egress optimization: Audio files are cached at edge nodes to minimize expensive inter-region
data transfer fees
• FinOps practices: Engineering teams have per-squad cloud cost dashboards for accountability
and optimization
3. Scalability
Horizontal Scaling
• GKE auto-scales microservice pods from 0 to thousands within seconds during peak events
(New Year, concert drops)
• Google Cloud's global load balancer distributes requests across regions in real-time
Vertical Scaling
• Database nodes (Bigtable, Cloud Spanner) scale storage and compute independently
• BigQuery serverless architecture scales query compute resources automatically
Database Scalability
• Cloud Spanner: Globally distributed, horizontally scalable relational database for user data with
99.999% SLA
• Bigtable: Used for real-time personalization data – scales to petabytes with millisecond latency
• Memorystore (Redis): Distributed in-memory cache for session data, playlist caching, and
search results
Event-Driven Scalability
• Google Pub/Sub handles billions of events per day (play events, skip events, search queries)
• Dataflow (Apache Beam) processes streaming data in real-time for recommendation updates
Page 7 of 14
Cloud Computing Assignment | March 2026
Content Delivery Scalability
• Audio files are pre-distributed across GCP's 100+ Points of Presence (PoPs) globally
• DASH (Dynamic Adaptive Streaming over HTTP) adjusts audio bitrate based on user's
connection speed
4. Security
Security Domain Techniques & Tools Used
Data Encryption AES-256 encryption at rest for all user data; TLS 1.3 for data in
transit between clients and servers
Authentication OAuth 2.0 for third-party app authorization; PKCE flow for mobile
clients; MFA for employee accounts via Okta
Zero Trust Network BeyondCorp-style zero-trust model: all internal service-to-service
calls require mutual TLS (mTLS) authentication
Secrets Management HashiCorp Vault for managing API keys, database credentials,
and service account tokens
DDoS Protection Google Cloud Armor with adaptive protection rules; Fastly WAF for
edge protection
Vulnerability Management Continuous scanning via Snyk and Dependabot; Security chaos
engineering (Game Day exercises)
Compliance GDPR, CCPA, SOC 2 Type II, ISO 27001 certified; regular third-
party penetration testing
Privacy Data minimization practices; user data anonymization for analytics;
right-to-erasure workflows
Incident Response 24/7 Security Operations Center (SOC); PagerDuty-based alerting;
automated runbooks in Confluence
Page 8 of 14
Cloud Computing Assignment | March 2026
Question 4: Cloud Migration Research – Methods, Types & IEEE
Reference
Overview of Cloud Migration
Cloud migration is the process of moving applications, data, and workloads from on-premises
infrastructure (or from one cloud) to a cloud environment. Recent research has focused on automated,
intelligent, and cost-optimized migration strategies.
Migration Methods & Techniques (from Recent Research Papers)
Method / Technique Description Research Focus
Rehost (Lift & Shift) Moving applications as-is to cloud Fastest migration; minimal risk;
VMs without code changes. Uses used in legacy enterprise systems
tools like AWS MGN, Azure Migrate,
Carbonite Migrate.
Replatform (Lift & Minor optimizations during Balances speed and cloud-native
Reshape) migration, such as moving to benefits; popular in SME
managed databases or migrations
containerizing applications.
Refactor / Re- Redesigning applications as Maximum cloud-native benefit;
architect microservices or serverless on intensive but high ROI
cloud. Involves decomposition using
Domain-Driven Design.
Repurchase Replacing existing software with a Reduces maintenance overhead;
(Replace) cloud-native SaaS product (e.g., suitable for commodity
moving CRM to Salesforce). applications
Retain (Revisit) Keeping certain applications on- Hybrid cloud strategy; common in
premises due to compliance, healthcare and banking
latency, or ROI constraints.
Retire Decommissioning applications no Cost reduction; portfolio
longer needed post-migration rationalization
analysis.
ML-based Workload Using machine learning to predict AI-driven migration planning;
Placement optimal cloud placement for reduces human effort in
workloads based on cost, latency, assessment
and compliance requirements.
Container-based Containerizing legacy apps with Popular in DevOps-mature
Migration Docker and migrating via organizations; enables multi-
Kubernetes. Uses tools like cloud portability
Konveyor Crane, Velero for state
migration.
Live Migration Moving VMs and workloads with Critical for high-availability
minimal downtime using hypervisor- systems; minimizes SLA
level migration or in-memory violations
replication.
Database Migration Schema conversion, data replication Most complex part of migration;
Page 9 of 14
Cloud Computing Assignment | March 2026
Method / Technique Description Research Focus
using AWS DMS, Google requires careful validation
Datastream, Striim. Handles
heterogeneous DB migrations.
Types of Cloud Migration
Type Description
On-Premises → Public Most common type. Moving from data centers to AWS, Azure, or
Cloud GCP. Reduces CAPEX to OPEX.
Public Cloud → Public Migrating between cloud providers for better pricing, features, or
Cloud (Cloud-to-Cloud) compliance. Also called cloud repatriation or lateral migration.
On-Premises → Private Building an internal cloud using OpenStack, VMware, or Nutanix for
Cloud compliance-sensitive workloads.
On-Premises → Hybrid Splitting workloads between on-premises and cloud. Common in
Cloud banking and healthcare sectors.
Single Cloud → Multi- Distributing workloads across multiple cloud providers to avoid vendor
Cloud lock-in and improve resilience.
Monolithic → Decomposing monolithic applications during migration using DDD,
Microservices Strangler Fig Pattern, or CQRS patterns.
(Application Migration)
Research Suggestions
• Focus on AI/ML-assisted migration assessment tools that auto-discover dependencies and
generate migration plans
• Develop standardized migration metrics: Total Cost of Migration (TCM), migration downtime,
data integrity score
• Research on reversible migrations (cloud repatriation) as organizations reassess cloud costs
post-migration
• Green cloud migration: optimizing migrations to reduce carbon footprint by choosing energy-
efficient regions
• Federated learning-based migration planning that shares insights without exposing sensitive
enterprise data
IEEE Reference Paper
Cited IEEE Paper:
M. Jamshidi, A. Sharifloo, C. Pahl, A. Metzger and G. Estrada, "Self-Learning Cloud Controllers:
Fuzzy Q-Learning for Knowledge Evolution," 2015 3rd International Conference on Cloud and
Autonomic Computing, IEEE, 2015, pp. 208-211. DOI: 10.1109/ICCAC.2015.24
Additional IEEE Reference:
Page 10 of 14
Cloud Computing Assignment | March 2026
T. Shet and N. Vadlamani, "A Survey on Cloud Migration Strategies," 2020 IEEE International
Conference on Electronics, Computing and Communication Technologies (CONECCT), IEEE,
2020, pp. 1-6. DOI: 10.1109/CONECCT50063.2020.9198439
Page 11 of 14
Cloud Computing Assignment | March 2026
Question 5: Cloud Security Research – Algorithms, Types &
Suggestions
Overview
Cloud security is the set of policies, controls, procedures, and technologies that protect cloud-based
systems, data, and infrastructure. Recent research has focused on AI-driven threat detection, zero-trust
architectures, and homomorphic encryption for privacy-preserving computation.
1. Algorithms & Techniques Used in Cloud Security (from Recent Papers)
Algorithm / Technique Security Purpose Details
AES-256 (Advanced Data Encryption Symmetric key algorithm. Industry standard for
Encryption Standard) encrypting data at rest in cloud storage (S3,
Azure Blob, GCS). 256-bit keys provide
quantum-resistant security margins.
RSA-2048 / RSA-4096 Key Exchange & Asymmetric encryption for TLS handshakes,
Authentication digital certificates, and PKI. Used in HTTPS,
SSH, and code signing across cloud platforms.
Elliptic Curve Lightweight Provides equivalent security to RSA with smaller
Cryptography (ECC) Encryption keys. Used in IoT-cloud communication, mobile
authentication, and TLS 1.3 cipher suites.
SHA-256 / SHA-3 Data Integrity Cryptographic hash functions used in cloud
(Hashing) storage checksums, digital signatures,
blockchain in cloud, and JWT token signing.
Homomorphic Privacy-Preserving Allows computation on encrypted data without
Encryption (HE) Computation decryption. Enables secure ML model training
on sensitive healthcare/financial cloud data.
(Microsoft SEAL, IBM HElib)
Zero-Knowledge Proofs Identity Verification Proves identity/ownership without revealing
(ZKP) secrets. Used in blockchain cloud platforms,
decentralized identity systems (DID), and
privacy-preserving audits.
Federated Learning Distributed ML Trains AI security models across distributed
Security cloud nodes without centralizing raw data.
Enables privacy-preserving threat detection
across organizations.
RBAC / ABAC (Access Authorization Role-Based and Attribute-Based Access
Control) Control. AWS IAM, Azure RBAC implement
these models to enforce least-privilege access
policies at scale.
Anomaly Detection (ML- Threat Detection Isolation Forest, LSTM autoencoders, and
based) Graph Neural Networks detect abnormal user
behavior, API call patterns, and insider threats in
cloud logs.
Deep Packet Inspection Network Security Analyzes cloud network traffic at Layer 7. Used
(DPI) in cloud firewalls (AWS Security Groups, Azure
Page 12 of 14
Cloud Computing Assignment | March 2026
Algorithm / Technique Security Purpose Details
NSG) to detect malware, exfiltration, and C2
communication.
Intrusion Detection with Security Monitoring Deep learning models classify network traffic as
CNN/RNN normal or attack. Trained on NSL-KDD, CICIDS
datasets. Deployed in cloud SIEM platforms.
Blockchain for Audit Immutable Logging Stores cloud access logs and configuration
Trails changes in distributed ledger. Prevents log
tampering. Used in regulatory compliance
(HIPAA, SOX).
2. Types of Cloud Security
Security Type Description Examples
Data Security Protecting data at rest, in transit, AWS KMS, Azure Key Vault,
and in use through encryption, Google Cloud DLP, Macie
masking, and DLP policies.
Identity & Access Controlling who can access what AWS IAM, Azure Active Directory,
Management (IAM) resources using authentication and Okta, Ping Identity
authorization policies.
Network Security Protecting cloud network perimeters, VPCs, Security Groups, NACLs,
segmenting VPCs, and monitoring AWS Shield, Azure Firewall
traffic.
Application Security Securing cloud-hosted applications WAF, SAST/DAST tools, Snyk,
(AppSec) against OWASP Top 10 Checkmarx, Veracode
vulnerabilities.
Endpoint Security Securing devices accessing cloud CrowdStrike Falcon, Microsoft
resources. Defender for Cloud, Jamf
Compliance & Ensuring cloud configurations meet AWS Config, Azure Policy, GCP
Governance regulatory standards. Security Command Center
Zero Trust Security Never trust, always verify – BeyondCorp, Zscaler, Palo Alto
continuous authentication for all Prisma Access
users and devices.
Container & Securing containerized workloads Falco, Aqua Security, Sysdig,
Kubernetes Security and orchestration platforms. Prisma Cloud
Incident Response & Detecting, responding to, and AWS GuardDuty, Azure Sentinel,
Forensics recovering from security incidents in Splunk SIEM, PagerDuty
cloud.
Physical Security Cloud provider secures physical Managed by AWS, Azure, GCP
data centers with biometrics, 24/7 (SOC 2 certified)
surveillance, and access controls.
3. Research Suggestions
• Post-Quantum Cryptography: Implement NIST-standardized algorithms (CRYSTALS-Kyber,
CRYSTALS-Dilithium) to future-proof cloud encryption against quantum attacks
Page 13 of 14
Cloud Computing Assignment | March 2026
• AI-native SOC: Develop autonomous Security Operations Centers using LLMs for alert triage,
threat hunting, and automated remediation
• Confidential Computing: Expand use of Intel SGX, AMD SEV, and AWS Nitro Enclaves for
secure processing of sensitive healthcare and financial data
• Supply Chain Security: Research SBOM (Software Bill of Materials) automation for cloud-native
applications to detect compromised dependencies early
• Unified Security Data Lake: Build open-standard (OCSF) security data lakes across multi-cloud
environments for correlated threat detection
IEEE Reference Paper
S. Subashini and V. Kavitha, "A survey on security issues in service delivery models of cloud
computing," Journal of Network and Computer Applications, Elsevier/IEEE, Vol. 34, Issue 1, 2011,
pp. 1-11. DOI: 10.1016/[Link].2010.07.006
Additional IEEE Reference:
X. Liu et al., "Cloud Security Using Blockchain and Machine Learning – A Comprehensive Survey,"
IEEE Access, Vol. 9, 2021, pp. 56956-56985. DOI: 10.1109/ACCESS.2021.3070927
Page 14 of 14