MyCDN - Perancangan Kapasiti On-Premises

Pendekatan Sizing: Workload-Driven vs Inventory-Driven

Dokumen ini menggunakan pendekatan Workload-Driven Sizing - infrastruktur diterbitkan daripada keperluan beban kerja sebenar, bukan sekadar senarai inventori perkakasan.

Mengapa Workload-Driven?

Workload-Driven (Pendekatan Kami)

Bermula dariPengguna, TPS, Transaksi

MetodologiBeban kerja menentukan infrastruktur

Soalan dijawab"Mengapa nombor ini?"

KelebihanBoleh disahkan, boleh diaudit

HasilSizing yang justified dan defensible

Inventory-Driven (Pendekatan Tradisional)

Bermula dariSenarai perkakasan

MetodologiInventori dahulu, justifikasi kemudian

Soalan dijawab"Apa yang ada?"

KelemahanSukar disahkan, over/under-sized

HasilSizing tanpa asas kukuh

Aliran Metodologi Kami

Step 1: Input Pengguna - Bermula dengan data pengguna sebenar daripada JKDM

Parameter Input	Nilai	Sumber
Jumlah Pengguna Berdaftar	644,500	Keperluan JKDM (pembayar cukai)
Pengguna Dalaman (JKDM Staff)	17,450	Keperluan JKDM
Peak Concurrency Rate	5%	Standard industri untuk sistem cukai
Peak Concurrent Users	~32,000	644,500 x 5%

Step 2: Derivasi TPS - Concurrent users diterjemahkan kepada transaksi

Parameter	Nilai	Derivasi
Average Session Duration	15-20 min	Typical tax filing session
Transactions per Session	3-5	Login, browse, submit, confirm, logout
Peak TPS (Backend)	500	Derived from concurrent users + transaction patterns
Frontend:Backend Ratio	10:1	10 frontend requests per backend transaction
Peak Frontend RPS	~5,000	500 TPS x 10

Step 3: Sizing Formula - TPS diterjemahkan kepada vCPU

Component	Formula	Hasil	Justifikasi
iTAX Core	TPS x vCPU/TPS x HA	128 vCPU	500 TPS x 0.12 vCPU/TPS x 2 (HA)
API Gateway	RPS x vCPU/RPS x HA	32 vCPU	5000 RPS x 0.003 vCPU/RPS x 2
ESB/Integration	Integrations x vCPU	48 vCPU	25+ integrations x 2 vCPU x buffer
Database	TPS x vCPU/TPS x RAC	128 vCPU	500 TPS x 0.13 vCPU/TPS x 2 (RAC)

Step 4: Degradation Model untuk DR - DRC bukan full mirror

Aspek	DC	DRC	Rasional
Capacity Target	100%	50%	Graceful degradation semasa failover
VMs	91	41	Critical workloads sahaja
vCPU	784	320	Essential processing capacity
Data Platform	Included	Excluded	Rebuild from source selepas failover

Graceful Degradation: Semasa failover, bukan semua pengguna memerlukan akses serentak pada peak capacity. DRC menyediakan kapasiti untuk operasi kritikal dengan reduced throughput, bukan full production mirror.

Headroom Philosophy

DC Headroom: 18%

Physical Capacity960 vCPU

Allocated784 vCPU

Headroom176 vCPU (18%)

PurposePeak loads, growth buffer

DRC Overcommit: 1.00x

Physical Capacity320 vCPU

Allocated320 vCPU

Overcommit Ratio1.00x (tiada)

PurposeFull capacity sedia semasa failover

Kesimpulan Metodologi: Setiap nombor dalam dokumen ini boleh di-trace balik kepada keperluan beban kerja. Tiada "magic numbers" - semua sizing boleh disahkan, diaudit, dan dipertahankan kepada panel penilai.

Sizing Termasuk 18% Operational Headroom

Sizing ini termasuk 18% headroom dalam baseline: DC 20 nodes (960 vCPU physical) dengan 784 allocated = 82% utilisation.

Headroom 176 vCPU menyediakan operational buffer untuk peak loads dan pertumbuhan jangka pendek tanpa perlu penambahan nod.

VM-Based Architecture: Semua komponen MyCDN di-deploy sebagai virtual machines (VMs), bukan containers. Tiada Kubernetes dalam skop ini. Semua services termasuk iTAX, API Gateway, ESB, Kafka, Redis dijalankan sebagai VMs.

HCI Servers

20 DC + 10 DRC

1,280

Physical vCPU

960 DC + 320 DRC

1,104

Allocated vCPU

784 DC + 320 DRC

132

Total VMs

91 DC + 41 DRC

82%

DC Utilisation

18% headroom

Definisi Terma Penting

Physical vCPU:Kapasiti fizikal HCI nodes (960 DC + 320 DRC = 1,280). Kapasiti maksimum sebelum overcommit.

Allocated vCPU:vCPU diperuntuk kepada VM (784 DC + 320 DRC = 1,104). DC 82% utilisation, DRC 1.00x (tiada overcommit).

Headroom:Buffer untuk peak loads dan growth. Headroom BUKAN untuk non-production (Dev/SIT/UAT).

Non-Production:Dev, SIT, UAT TIDAK TERMASUK dalam sizing ini. Akan diperuntuk secara berasingan.

Ringkasan Infrastruktur On-Premises (Reconciled)

Komponen	DC	DRC	Jumlah	Nota
HCI Server Nodes	20	10	30	Dell VxRail / HPE SimpliVity
Physical vCPU	960	320	1,280	Kapasiti fizikal HCI
Allocated vCPU (PROD)	680	320	1,000	Production workloads
Allocated vCPU (Data Platform)	104	-	104	Analytics/BI (DC sahaja)
Total Allocated vCPU	784	320	1,104	PROD + Data Platform
Utilisation / Overcommit	82%	1.00x	-	DC: 784/960, DRC: 320/320
Headroom	176 vCPU (18%)	-	-	Operational buffer untuk peak loads
RAM	20,480 GB	7,680 GB	28,160 GB	~28TB total
Primary Storage	176 TB	65 TB	241 TB	HCI + SAN + Object + DW
VMs (PROD)	82	41	123	DRC = 50% VM count
VMs (Data Platform)	9	-	9	DC sahaja, tidak replicated
Total VMs	91	41	132	Semua VM-based (Linux)

Sizing termasuk 18% headroom: DC 20 nodes menyediakan 960 vCPU physical dengan 784 allocated = 82% utilisation. Headroom 176 vCPU (18%) untuk operational buffer dan peak loads. DRC beroperasi pada 1.00x (tiada overcommit).

Gambarajah Arsitektur: Bahagian ini menunjukkan rekabentuk infrastruktur on-premises MyCDN termasuk topologi DC-DRC, komponen keselamatan DMZ, aplikasi iTAX/MyCDN, Data Platform, dan aliran data. 100% on-premises dengan VMware vSphere 8 + vSAN.

Seni Bina Infrastruktur MyCDN (On-Premises)

Nota Seni Bina: Diagram menunjukkan topologi lengkap DC dan DRC di Kelana Jaya dengan komponen keselamatan DMZ, aplikasi iTAX/MyCDN, Data Platform (Greenplum, Tableau, Object Storage), dan pangkalan data PostgreSQL HA. Tiada komponen cloud - 100% on-premises.

3. VM Distribution Summary

DC VMs by Layer91 VMs

iTAX Core8 VMs (128 vCPU)

iTAX Database4 VMs (128 vCPU)

MyCDN Custom12 VMs (96 vCPU)

API Gateway4 VMs (32 vCPU)

ESB / Integration6 VMs (48 vCPU)

Kafka + Redis10 VMs (64 vCPU)

Supporting (Security, Monitor, etc)38 VMs (184 vCPU)

Data Platform9 VMs (104 vCPU)

Total DC91 VMs (784 vCPU)

DRC VMs (50% Replica)41 VMs

iTAX Core4 VMs (64 vCPU)

iTAX Database2 VMs (64 vCPU)

MyCDN Custom6 VMs (48 vCPU)

API Gateway2 VMs (16 vCPU)

ESB / Integration3 VMs (24 vCPU)

Kafka + Redis5 VMs (32 vCPU)

Supporting19 VMs (72 vCPU)

Data PlatformNot Replicated

Total DRC41 VMs (320 vCPU)

DC Sizing - Dengan 18% Operational Headroom

DC menyediakan 18% headroom: 960 physical vCPU (20 nodes) dengan 784 allocated = 82% utilisation.

Headroom 176 vCPU membolehkan operational buffer untuk peak loads dan growth tanpa penambahan nod.

960

Physical vCPU

20 nodes x 48

784

Allocated vCPU

680 PROD + 104 Data

82%

Utilisation

18% headroom

Total VMs

82 PROD + 9 Data

DC - Production Components (82 VMs, 680 vCPU)

Component	VM	vCPU/VM	Total vCPU	RAM/VM	Total RAM
iTAX Core Application	8	16	128	64GB	512GB
iTAX Database (Oracle)	4	32	128	128GB	512GB
MyCDN Custom Modules	12	8	96	32GB	384GB
API Gateway	4	8	32	16GB	64GB
ESB / Integration Hub	6	8	48	32GB	192GB
Message Broker (Kafka)	6	8	48	32GB	192GB
Cache Layer (Redis)	4	4	16	32GB	128GB
Monitoring Stack	6	4	24	16GB	96GB
Load Balancer	4	4	16	8GB	32GB
Security Services	6	4	24	16GB	96GB
Identity Management	4	4	16	16GB	64GB
Backup & Recovery	4	4	16	16GB	64GB
Management & Ops	14	6	88	16GB	224GB
PROD SUBTOTAL	82	-	680	-	2,560GB

DC - Data Platform Components (9 VMs, 104 vCPU)

Component	VM	vCPU/VM	Total vCPU	RAM/VM	Total RAM
Data Warehouse (Greenplum)	4	16	64	64GB	256GB
ETL Processing	3	8	24	32GB	96GB
BI/Reporting (Tableau)	2	8	16	32GB	64GB
DATA PLATFORM SUBTOTAL	9	-	104	-	416GB

Pengesahan Matematik: DC Physical = 960 (20 nodes x 48). DC Allocated = 680 (PROD) + 104 (Data Platform) = 784. Utilisation = 784/960 = 82%. Headroom = 176 vCPU (18%).

DRC Sizing - Penjelasan Overcommit Logic

DRC mengandaikan partial workload activation semasa failover, bukan full concurrent peak.

Apabila failover berlaku, bukan semua workload diaktifkan serentak pada peak capacity. Pengguna dialihkan secara berperingkat (graceful degradation).

DRC beroperasi pada 1.00x (tiada overcommit) memastikan kapasiti penuh tersedia semasa failover.

320

Physical vCPU

10 nodes x 32

320

Allocated vCPU

PROD replicas

1.00x

Overcommit

320/320

Total VMs

50% of DC PROD

Penjelasan: DRC VM Count adalah Logical Mirror

DRC VM count (41) adalah 50% daripada DC PROD VM count (82). Ini adalah logical mirror, bukan full runtime mirror.

DRC menjalankan VM yang sama dari segi fungsi, tetapi dengan reduced count. Tidak semua instances aktif serentak.

DRC Component Breakdown (41 VMs, 320 vCPU)

Component	DC VM	DRC VM	vCPU/VM	DRC vCPU	Notes
iTAX Core Application	8	4	16	64	50% replica
iTAX Database (Oracle)	4	2	32	64	Standby only
MyCDN Custom Modules	12	6	8	48	Critical modules
API Gateway	4	2	8	16	50% replica
ESB / Integration Hub	6	3	8	24	Critical integrations
Message Broker (Kafka)	6	3	8	24	Minimum quorum
Cache Layer (Redis)	4	2	4	8	Warm standby
Monitoring Stack	6	3	4	12	Basic monitoring
Load Balancer	4	2	4	8	Active-passive
Security Services	6	3	4	12	Essential security
Identity Management	4	2	4	8	Replicated
Backup & Recovery	4	2	4	8	DR backup
Management & Ops	14	7	4	24	Basic management
DRC TOTAL	82	41	-	320	50% VM, 47% vCPU

DRC Failover Strategy: Hot-warm standby. Database replicas sentiasa synchronised. Application VMs dalam warm state. Graceful degradation: non-critical workloads dikurangkan.

Data Platform adalah DC-Only - Tidak Replicated ke DRC

Data Platform (Analytics/BI) tidak di-replicate ke DRC. Jika failover, Data Platform di-rebuild dari transactional sources.

Implikasi: Data Platform capacity (104 vCPU, ~41TB storage) untuk DC sahaja. DRC storage jauh lebih rendah kerana Data Platform tidak replicated.

Data Platform Components (DC Only)

Component	Location	vCPU	RAM	Storage	Replicated?
Data Warehouse (Greenplum)	DC Only	64	256GB	10TB	No - rebuild
Data Lake / Object Storage	DC Only	-	-	30TB	No - rebuild
ETL Processing	DC Only	24	96GB	500GB	No - stateless
BI/Reporting (Tableau)	DC Only	16	64GB	500GB	No - config only
DATA PLATFORM TOTAL	DC Only	104	416GB	~41TB	-

Storage Comparison: Kenapa DRC Storage Lebih Rendah?

DC Storage160TB

HCI Storage (VMs)70TB

SAN Storage (Database)50TB

Object Storage (Data Lake)30TB

Data Warehouse10TB

Total DC160TB

DRC Storage65TB

HCI Storage (VMs)30TB

SAN Storage (Database)35TB

Object Storage0TB (not replicated)

Data Warehouse0TB (not replicated)

Total DRC65TB

Jawapan "Kenapa DR storage rendah?": DRC (65TB) = ~41% DC (160TB) kerana Data Platform (~40TB) tidak replicated. Keputusan arsitektur yang disengajakan.

Capacity Verification - Matematik Pengesahan

Check	DC Allocated	DC Physical	DRC Allocated	DRC Physical	Status
vCPU (PROD)	680	960	320	320	OK
vCPU (Data Platform)	104	-	-	-	OK (DC only)
vCPU Total	784	-	320	-	DC 82%, DRC 1.00x
DC Utilisation	784 / 960 = 82%	-	320 / 320 = 100%	-	DC OK with 18% headroom
DC Headroom	176 vCPU (18%)	-	-	-	Operational buffer
VM Count (PROD)	82	-	41	-	DRC = 50% DC
VM Count (Data Platform)	9	-	0	-	DC only
Total VMs	91	-	41	-	132 total

Formula Verification

DC Calculation

Physical vCPU20 nodes x 48 = 960

PROD Allocated680 vCPU

Data Platform Allocated104 vCPU

Total Allocated680 + 104 = 784

Utilisation784 / 960 = 82%

Headroom176 vCPU (18%)

DRC Calculation

Physical vCPU10 nodes x 32 = 320

PROD Allocated320 vCPU

Data Platform0 (not replicated)

Total Allocated320

Overcommit Ratio320 / 320 = 1.00x

StatusOK (No Overcommit)

Verification Complete: Semua nombor konsisten. DC = 82% utilisation dengan 18% headroom. DRC = 1.00x (tiada overcommit).

Pernyataan Arsitektur Utama: iTAX merupakan aplikasi monolitik berasaskan .NET (runtime lintas-platform) dengan frontend Angular, dideploy pada Linux VM (RHEL 8/9). Runtime .NET lintas-platform menyokong Linux, membolehkan standardisasi OS dan pengurangan kos lesen. Komponen-komponen lain adalah modul sokongan, bukan microservices bebas.

Clarification Statements (Untuk Ketelusan Teknikal)

1. Core Application: iTAX adalah aplikasi monolitik berasaskan .NET (runtime lintas-platform) dengan Angular frontend, di-host pada Linux (RHEL 8/9) VMs. Runtime .NET lintas-platform menyokong deployment pada Linux tanpa kebergantungan Windows. Ini bukan cloud-native atau container-based architecture.

2. Service Components: Komponen yang disenaraikan adalah logical application modules, bukan container-based microservices. Tiada service discovery, independent scaling, atau CI/CD isolation per-service.

3. Messaging & Caching: Kafka dan Redis adalah komponen sokongan untuk reliability dan async processing, bukan primary transaction engines. Core transaction flow adalah synchronous. Kafka dipilih berbanding MSMQ kerana keperluan high-volume audit logging, cross-platform notification dispatch, dan guaranteed delivery untuk downstream integrations.

4. CRM & DMS: Sistem CRM dan DMS adalah auxiliary enterprise systems yang diintegrasikan dengan MyCDN, bukan core processing workloads. CRM dan DMS diandaikan sebagai sistem sedia ada atau diperuntukkan secara berasingan. Hanya integration endpoints yang termasuk dalam sizing MyCDN.

5. Network Security: Keselamatan infrastruktur dikuatkuasakan di network dan perimeter layers. NSX microsegmentation adalah optional enhancement, bukan keperluan baseline.

Core Application Architecture

Aspect	Specification	Implication
Runtime	.NET (lintas-platform / cross-platform)	Linux (RHEL 8/9), tiada kebergantungan Windows
Architecture Pattern	Monolithic with modular components	Not microservices, not independently scalable
Frontend	Angular SPA	Served from Nginx on Linux
Scaling Model	Vertical + horizontal VM cloning	Not elastic auto-scaling
Deployment Unit	Linux VM (RHEL 8/9)	Full application per VM, not container pods
Session Management	Stateful with session affinity	Load balancer sticky sessions required

Summary: This is a VM-based, monolithic COTS application built on .NET (cross-platform runtime) with supporting components for integration, caching, and messaging. The .NET cross-platform runtime enables Linux deployment without Windows dependency, allowing OS standardisation and reduced licensing costs. It is not a cloud-native microservices architecture. This is appropriate for a COTS-based government system requiring stability and vendor support.

VMware Infrastructure Stack

Layer	Technology	Version	Required?	Notes
Hypervisor	VMware vSphere ESXi	8.x	Required	Enterprise Plus license
Management	VMware vCenter Server	8.x	Required	Centralized VM management
Storage	VMware vSAN	8.x	Required	HCI storage layer
Network Virtualization	VMware NSX	4.x	Optional	Microsegmentation optional, not baseline
DR Replication	VMware SRM	8.x	Required	DC to DRC replication
Backup	Veeam Backup & Replication	12.x	Required	VM-level backup

NSX microsegmentation provides enhanced east-west security but is not required for baseline operation. Network security is enforced at perimeter (firewall, WAF) and VLAN segmentation.

Application Components (All VM-Based, Not Microservices)

Component	Technology	Type	Role	Justification
iTAX Core	.NET (lintas-platform)	Core Engine	Tax processing	COTS licensed from Qualysoft, .NET runtime menyokong Linux
iTAX Database	Oracle 19c	Core Engine	Transaction store	Primary OLTP database
MyCDN Custom Modules	Java	Logical Module	Custom extensions	Malaysia-specific business logic
API Gateway	Kong Gateway	Supporting	External API exposure	Rate limiting, authentication for external consumers
ESB	WSO2 EI	Supporting	Integration hub	Legacy system adapters, protocol transformation
Message Broker	Apache Kafka	Supporting	Async decoupling	High-volume audit logging, notification dispatch, decoupled downstream processing with guaranteed delivery
Cache	Redis	Supporting	Performance	Session cache, reference data cache, not primary store
Data Warehouse	Greenplum	Analytics	BI/Reporting	Historical analysis, not real-time
BI/Reporting	Tableau Server	Analytics	Dashboards	Management reporting

Auxiliary Enterprise Systems (Not Core Processing)

System	Role	Integration Type	Sizing Impact
CRM	Customer relationship management	API integration with iTAX	Not in TPS model - low volume
DMS	Document management	File storage integration	Storage only - not in compute TPS

CRM and DMS are assumed to be existing or separately provisioned enterprise systems. Only integration endpoints are included in MyCDN sizing. No compute or storage allocation for CRM/DMS is included in this infrastructure.

User Concurrency Model

Parameter	Value	Derivation
Total Registered Users	644,500	From JKDM requirements
Peak Concurrent User %	5%	Industry standard for tax systems
Peak Concurrent Users	~32,000	644,500 x 5%
Average Session Duration	15-20 minutes	Tax filing typical session
Peak TPS	500	Based on concurrent users and transaction patterns
FE:BE Ratio	~10:1	10 frontend requests per backend transaction

Sizing Calculator: Kalkulator ini membolehkan panel penilai mengubah parameter input dan melihat kesan kepada sizing secara automatik. Semua konstanta sizing dinyatakan untuk ketelusan.

A. Input Parameters (Editable)

User & Traffic

1. Total Registered Users

Jumlah pengguna berdaftar sistem

2. Peak Concurrent User %

% pengguna serentak semasa peak (3-10%)

3. Peak TPS (Transactions/sec)

Target peak transactions per second

Operational Buffer & DR Policy

4. Operational Headroom (%)

Buffer untuk peak (baseline = 18%)

5. DR Capacity Factor

DRC as % of DC (0.5 = 50%)

6. DR Overcommit Ratio

DRC overcommit (1.00 = none)

Non-Production Environments (5 Persekitaran MyCDN)

7. Dev (% of PROD)

Development (baseline = 15%)

8. SIT (% of PROD)

System Integration Test (baseline = 25%)

9. UAT (% of PROD)

User Acceptance Test (baseline = 50%)

10. Staging (% of PROD)

Pre-Production (baseline = 75%)

B. Production Outputs (Auto-Updated)

32,225

Peak Concurrent Users

784

DC vCPU Allocated

DC HCI Nodes

320

DRC vCPU Allocated

DRC HCI Nodes

PROD HCI Nodes

176

Headroom vCPU

82%

DC Utilisation

C. Non-Production Outputs (Auto-Updated)

102

Dev vCPU (15%)

Dev HCI Nodes

170

SIT vCPU (25%)

SIT HCI Nodes

340

UAT vCPU (50%)

UAT HCI Nodes

510

Staging vCPU (75%)

Staging HCI Nodes

Non-Prod HCI Nodes

GRAND TOTAL Nodes

D. Sizing Constants & Assumptions

Baseline sizing derived from detailed component analysis. Scaling is linear from baseline:

Constant	Value	Description
Baseline TPS	500 TPS	Reference point for sizing calculations
Baseline PROD vCPU	680 vCPU	PROD vCPU required at 500 TPS (from component breakdown)
Data Platform vCPU	104 vCPU	Fixed allocation for analytics/BI (DC only)
DC vCPU per Node	48 vCPU	Usable vCPU per DC HCI node (dual-socket)
DRC vCPU per Node	32 vCPU	Usable vCPU per DRC HCI node
Min DC Nodes	12	Minimum DC nodes for HA cluster
Min DRC Nodes	6	Minimum DRC nodes for HA
Scaling Formula	Linear	PROD vCPU = 680 x (TPS / 500)

Note: At default settings (500 TPS, 50% DR, 1.00x overcommit, 18% headroom), calculator output matches document baseline: 30 PROD nodes. Non-Production adds UAT/SIT/Dev environments.

E. What-If Scenarios (Pre-Calculated)

Below shows sizing impact at different TPS levels (18% headroom included, 50% DR factor, 1.00x DRC):

Scenario	TPS	PROD vCPU	DC vCPU	DC Nodes	DRC vCPU	DRC Nodes	Total
Current Baseline	500	680	784	20	320	10	30
+20% Growth	600	816	920	20	408	14	34
+40% Growth	700	952	1,056	22	476	16	38
+60% Growth	800	1,088	1,192	26	544	18	44

Note: Baseline includes 18% operational headroom. DC utilisation at baseline = 82%.

Calculator Purpose: This calculator provides indicative estimates for what-if analysis. Change TPS, headroom, or DR parameters to see projected infrastructure requirements. At default settings, output matches document baseline (30 nodes with 18% headroom).

Untuk Panel Penilai: Tab ini menyediakan ringkasan mudah dan analogi untuk membantu memahami sizing infrastruktur.

Analogi Mudah

Bayangkan MyCDN seperti Hospital Besar

DC (Data Centre):Hospital utama beroperasi 24/7 dengan semua kemudahan lengkap.

DRC (DR Centre):Hospital sandaran - beroperasi penuh jika hospital utama ditutup. Biasanya "siap sedia".

vCPU:Bilangan doktor. Lebih banyak = lebih ramai pesakit boleh dilayan serentak.

Headroom:Doktor tambahan untuk waktu kecemasan. Jika tiada, hospital sesak semasa peak.

Overcommit DRC:Hospital sandaran ada 10% kurang doktor - boleh diterima kerana pesakit tidak datang serentak.

Soalan Lazim (FAQ)

Soalan	Jawapan
Adakah termasuk headroom?	Ya. DC 20 nodes = 960 vCPU physical, 784 allocated = 82% utilisation. 18% headroom untuk operational buffer.
Kenapa DRC storage rendah?	Data Platform (~40TB analytics) tidak replicate ke DRC. DRC hanya untuk PROD data.
Kenapa DRC overcommit 10%?	DRC mengandaikan partial activation semasa failover, bukan 100% serentak.
Dev/SIT/UAT di mana?	Tidak termasuk dalam sizing ini. Non-prod diperuntuk secara berasingan.
Adakah gunakan Kubernetes?	Tidak. Semua komponen adalah VM-based menggunakan VMware stack.
Jika TPS naik ke 700?	Guna Calculator tab. Anggaran: perlu +4 DC nodes, +4 DRC nodes.

Ringkasan Nombor Utama

HCI Servers

1,104

Allocated vCPU

132

Total VMs

241TB

Total Storage