AWS Cloud Infrastructure Optimization and Modernization

Customer Context

dugun.com is a premier digital platform and marketplace dedicated to the wedding industry, serving as a
comprehensive planning ecosystem that connects millions of couples with thousands of specialized vendors
and event professionals. Managing large volumes of interactive traffic, digital assets, and high-velocity
database requests, the platform relies heavily on cloud infrastructure to maintain persistent availability, low-
latency search functionalities, and data analytics.

Starting Point

Prior to the optimization engagement, the client was already running its core applications entirely on AWS.
However, the ecosystem was impacted by architectural patterns that limited operational efficiency, bloated
cloud expenditures, and introduced security risks:

Legacy Compute & Network Architecture:

Application workloads were distributed across older-
generation compute instances. A significant security and management overhead existed because the
majority of these servers were provisioned with public IP addresses within public subnets, exposing them
directly to the internet.

Outdated Database Infrastructure:

Database layers were running on legacy database engine versions.
Critical analytical workloads relied on standard cluster configurations that suffered from severe I/O
inefficiencies and high operational overhead. Core transactional databases were bound to provisioned
IOPS storage types that were over-provisioned for actual utilization.

Idle and Orphaned Assets:

Lack of regular environment cleanup resulted in multiple untracked and idle
assets, including unattached block storage volumes, deprecated observability dashboards, unused
database proxies, and unassigned static IP addresses that incurred ongoing idle charges.

Target Architecture on AWS

Compute & Network Security Hardening

The entire server fleet was modernized by upgrading legacy workloads to current-generation, cost-optimized
compute instance families, significantly enhancing performance per watt and core processing efficiency. To
establish a robust security baseline, instances were migrated from exposed public subnets into isolated
private subnets. External access was securely routed through Application Load Balancers (ALBs) and highly
available NAT Gateways, effectively minimizing the public attack surface and eliminating the need for
excessive public IP allocations and idle static IPs.

Storage & Database Modernization

The database environment underwent a substantial technical transformation to optimize costs and version
compatibility:

Database Engine Upgrades:

Core transactional and analytical databases were upgraded to modernized,
fully supported engine versions to ensure compatibility, security patches, and performance improvements.

Storage Tier Optimization:

High-I/O analytical database clusters were migrated from standard models to
advanced I/O-optimized storage architectures, completely mitigating excessive input/output transactional
costs. Unused database proxies were completely decommissioned.

Block Storage Standardization: Legacy storage structures and provisioned IOPS volumes were
seamlessly converted to modern general-purpose block storage volumes ( gp3 ) across both database
clusters and standalone instances, delivering baseline high performance without provisioned cost
penalties. Any unattached, orphaned block storage volumes were purged.

Object Storage & Network Routing:

Automated object storage lifecycle policies were introduced to
automatically transition or archive aging data. Amazon CloudFront was integrated in front of public storage
buckets to cache assets globally and decrease external data transfer fees. Furthermore, a Gateway VPC
Endpoint for object storage was established, keeping inter-service traffic entirely within the private AWS
network topology to bypass external data processing charges.

Observability & Network Management

The monitoring layer was refined by purging legacy, inactive dashboards within Amazon CloudWatch ,
establishing high-signal alerting thresholds for current resource metrics. Network connectivity was simplified
by deprecating commercial licensing models for secure endpoints and adopting modern, highly efficient, and
open-source networking alternatives to handle backend admin access securely.

Business / Process Impact

The comprehensive transformation provided measurable business value and operational maturity:

Immediate Bottom-Line Savings:

Eliminating architectural inefficiencies, transitioning to modern
storage configurations, and purging unattached resources resulted in a massive immediate reduction
in monthly infrastructure run-rates without reducing operational capacity.

Hardened Security Posture:

Re-architecting the network topology to place compute resources into
private subnets eliminated exposure risks, drastically decreasing the platform’s external attack
surface.

Long-Term Predictability:

SStandardizing instance types enabled the organization to adopt
structured, multi-option annual commitment strategies (such as No Upfront or Partial Upfront
reservations for compute and caching layers), locking in prolonged cost stability.

Streamlined Infrastructure Operations:

Decommissioning legacy dashboards, orphaned disks, and
outdated proxies reduced operational clutter, leaving a clean, highly observable, and secure
production environment.