Cisco HyperFlex Sizing Report

Scenario: Udemy_Test

Customer: Udemy

Generated by: admin

Date: 2021-03-11
Disclaimer

The Hyperflex Sizer tool is designed to provide general guidance in evaluating the optimum solution for
using selected Cisco products, services and/or solutions. The tool is not intended to provide business,
legal, accounting, tax or professional advice. The tool is not intended as a substitute for User's own
judgment or for that of User's own professional advisors. The tool may not contain complete, accurate or
timely data. (For example, reference tables and other information embedded in the tool may require
occasional updating, and may not be up-to-date.) Use of the tool may not be error-free. While Cisco, its
licensors or suppliers have attempted to include useful data, assumptions and formulae in the tool, Cisco,
its licensors or suppliers are not responsible for any results from using or misusing the tool.
Use of the Tool is subject to our cisco.com Site terms at: https://www.cisco.com/c/en/us/about/legal/terms-
conditions.html
IN ADDITION, THE TOOL IS PROVIDED "AS IS," WITH NO WARRANTIES OF ANY KIND, INCLUDING
BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, NON-INFRINGEMENT OR
FITNESS FOR A PARTICULAR PURPOSE. ANY USE OF THE TOOL IS AT THE USER'S SOLE RISK. DO
NOT USE THE TOOL UNLESS YOU AGREE TO THESE TERMS AND CONDITIONS.

© 2020 Cisco and/or its affiliates. All rights reserved.

Content

▪ HyperFlex Product Overvie

▪ User Input
▪ Overall Sizing Result
▪ Lowest_Cost Sizing Repor
▪ All-Flash Sizing Repor
▪ All NVMe Sizing Repor
▪ Fixed Config Sizing Repor
▪ Workload Summary

© 2020 Cisco and/or its affiliates. All rights reserved.

s

Cisco HyperFlex
Overview
 One Architecture for Operational Simplicity
Unified Computing System
UCS Management Intersight IMC Supervisor

VM VM VM

HyperFlex UCS UCS

UCS Mini Fourth Gen. UCS Systems C240 C3000

Mainstream Converged Hyperconverged Software-Defined Scale Out

Computing Infrastructure Infrastructure Storage

EDGE CORE DATA CENTER CLOUD

© 2020 Cisco and/or its affiliates. All rights reserved.

 Datacenter Modernization
Your bridge to the Future
Optimized for the
Traditional Modernized Distributed
Infrastructure Hyperconverged Datacenter
Highly Manual Core Operation Infrastructur
Limited Time and Budget Streamlined and Modernized
for Future Initiatives Core

Core Cloud
63% of Organizations
Currently Describe Their
Current Workload at
“Highly Manual” Edge

Management
451 RESEARCH, 75% Time Savings3
VOICE OF THE ENTERPRISE
Downtime
90% Reduction7
Savings vs. 3-tier
80% Infrastructure1

© 2020 Cisco and/or its affiliates. All rights reserved.

e

What is HyperFlex?
Engineered and Optimized HCI

Performance and Scalability to

Powe
Any App, Any Cloud, Anywhere

HX Data Platform Cloud Platform UCS Foundations

• Purpose Built for HCI • Cisco Intersight • Software Defined Compute
• Fast and Flexible • Cisco CloudCenter • Fabric Centric
• Enterprise Grade Storage • Cisco Workload Optimization Mgr. • Intent-Based

© 2020 Cisco and/or its affiliates. All rights reserved.

r

 HCI with Converged Fabric Networking

H X 2 2 0 C
+
Highest Performanc
(Performance Critical)

Network
1
Integrated HCI
4 0 / 2 5 / 1 0 G B P

+
U C S F A B R I C N E T W O R K I N G
Policy Based 90+ Nodes, Single Fabric,
2 Single Management Domain
Management H X 2 2 0 C H X 2 4 0 C
High Performance in Small Footprint High Density with Performanc
(Databases, VDI, VSI) (Databases, VDI, VSI, Test/Dev)
End-to-End
3
HCI Automation

H X 2 2 0 C & E D G E H X 2 4 0 C
+
Smallest Footprin Capacity-Heav
(VDI, VSI, ROBO) (VDI, VSI, Test/Dev)

HYBRID CLUSTER ALL FLASH CLUSTER ALL NVME CLUSTER

© 2020 Cisco and/or its affiliates. All rights reserved.

y

 HyperFlex Feature Overview

#1 Performanc Data Managemen Seamless Scal Next-Gen HC Cloud Integration

Industry leading data- Inline storage efficiency, Edge to Datacenter, End to end automation, Multicloud ready, Intersight
distributed architecture snapshots & clones scale-out performance, including networking for mgmt,analytics,support
capacity, & compute

Securit HW Assiste Data Protectio Ease of Us Global Suppor

SED, PSIRT, Cisco Security Co-engineered HW + SW, Native replication, disaster Simplicity of eval, sizing, Global reach, Awarding
Development Lifecycle flexible HW configurations recovery & backup solns Day 0/1, support workflows winning Cisco TAC

© 2020 Cisco and/or its affiliates. All rights reserved.

y

Sizing Report

© 2018 Cisco and/or its affiliates. All rights reserved.

User Inputs
Workload Summary Sizing Configurations
Fault Configuration Description
Workload Type Description Count RF
Tolerance
HXDP Version 4.5
VSI-1 VSI Medium 60 3 N+1
Sizer Version 10.2.4

Cluster Type Hyperflex and Compute

Threshold Standard

Hypervisor ESXi

Sizing Option Bundle and CTO

Disk Option All Disks

Cache Option All Cache Disks

Software Cost 3 Year(s)

Hardware
Auto
Acceleration

HyperFlex Boost Auto

Single Cluster No

CPU Generation Recommended

Click here to know more on the best practices needed for the workloads

© 2020 Cisco and/or its affiliates. All rights reserved.

Lowest_Cost Sizing Report

© 2018 Cisco and/or its affiliates. All rights reserved.

Assignment of Workload to Clusters (1/1)

Number of Clusters Created: 1

Workload Type Description Count RF Fault Tolerance Cluster Assignment

VSI-1 VSI Medium 60 3 N+1 Cluster 1

© 2020 Cisco and/or its affiliates. All rights reserved.

Overall Sizing Results

Sizing Summary Recommended Nodes (1/1)

Number of Clusters 1 HXAF-SP-240M5SXC-V [HX-STD-04B] [Cluster 1]

2xIntel Xeon Gold 5220R Processor, 24 cores, 2.2 GHz | 768 [12x64] GiB DDR4 RAM (8 Base Sticks, 4 Add-On) | 3
Number of Nodes 2 + 1 (FT)
6x3.84TB, 2.5" SSD | 1x800GB SAS | 2 RU

Number of RU 6

© 2020 Cisco and/or its affiliates. All rights reserved.

 Cluster 1 Overall Sizing Summary
Expected cluster utilization based on total resources required by all workloads and total available resources on this cluster
CPU RAM

Total Usable* CPUs = 446 vCPUs Total Usable* RAM = 1,392 GiB
(After System overhead, CPU overprovisioning ratio = 4, (After System overhead, RAM overprovisioning ratio = 1,
Performance headroom = Sustain 1 node failure) Performance headroom = Sustain 1 node failure)

Workload Unused Workload Unused

PU 240 vCPUs 205 vCPUs RAM 960 GiB 432 GiB

0 111
Best practice* sizing threshold: 401 vCPUs 223 334 445 0 348 696
Best practice* sizing threshold: 1,280.6 GiB1044 1392

Storage Capacity Storage IOPS

Total Usable* Capacity = 21.2 TB (19.3 TiB) Total available IOPS = 57K
Total Effective* Capacity = 29.4 TB (26.8 TiB) Workload Model: 8K IO Block Size, 70/30 Read/Write
(After Replication Factor = 3, System overhead,
Effective storage savings= 38.9%)
Workload Free Space Available Workload Unused / Performance Headroom
city 14 TB 15 TB Storage IOPS 10 % 90 %

0 7 15 22 29
Best practice* sizing threshold: 20.6 TB (18.7 TiB) of Effective Capacity 0 25 50 75 100

© 2020 Cisco and/or its affiliates. All rights reserved.

* Definitions for Usable, Effective and Best Practice defined in the Glossary in the Appendix.
All-Flash Sizing Report

© 2018 Cisco and/or its affiliates. All rights reserved.

Assignment of Workload to Clusters (1/1)

Number of Clusters Created: 1

Workload Type Description Count RF Fault Tolerance Cluster Assignment

VSI-1 VSI Medium 60 3 N+1 Cluster 1

© 2020 Cisco and/or its affiliates. All rights reserved.

Overall Sizing Results

Sizing Summary Recommended Nodes (1/1)

Number of Clusters 1 HXAF-SP-240M5SXC-V [HX-STD-04B] [Cluster 1]

2xIntel Xeon Gold 5220R Processor, 24 cores, 2.2 GHz | 768 [12x64] GiB DDR4 RAM (8 Base Sticks, 4 Add-On) | 3
Number of Nodes 2 + 1 (FT)
6x3.84TB, 2.5" SSD | 1x800GB SAS | 2 RU

Number of RU 6

© 2020 Cisco and/or its affiliates. All rights reserved.

 Cluster 1 Overall Sizing Summary
Expected cluster utilization based on total resources required by all workloads and total available resources on this cluster
CPU RAM

Total Usable* CPUs = 446 vCPUs Total Usable* RAM = 1,392 GiB
(After System overhead, CPU overprovisioning ratio = 4, (After System overhead, RAM overprovisioning ratio = 1,
Performance headroom = Sustain 1 node failure) Performance headroom = Sustain 1 node failure)

Workload Unused Workload Unused

PU 240 vCPUs 205 vCPUs RAM 960 GiB 432 GiB

0 111
Best practice* sizing threshold: 401 vCPUs 223 334 445 0 348 696
Best practice* sizing threshold: 1,280.6 GiB1044 1392

Storage Capacity Storage IOPS

Total Usable* Capacity = 21.2 TB (19.3 TiB) Total available IOPS = 57K
Total Effective* Capacity = 29.4 TB (26.8 TiB) Workload Model: 8K IO Block Size, 70/30 Read/Write
(After Replication Factor = 3, System overhead,
Effective storage savings= 38.9%)
Workload Free Space Available Workload Unused / Performance Headroom
city 14 TB 15 TB Storage IOPS 10 % 90 %

0 7 15 22 29
Best practice* sizing threshold: 20.6 TB (18.7 TiB) of Effective Capacity 0 25 50 75 100

© 2020 Cisco and/or its affiliates. All rights reserved.

* Definitions for Usable, Effective and Best Practice defined in the Glossary in the Appendix.
All NVMe Sizing Report

© 2018 Cisco and/or its affiliates. All rights reserved.

Assignment of Workload to Clusters (1/1)

Number of Clusters Created: 1

Workload Type Description Count RF Fault Tolerance Cluster Assignment

VSI-1 VSI Medium 60 3 N+1 Cluster 1

© 2020 Cisco and/or its affiliates. All rights reserved.

Overall Sizing Results

Sizing Summary Recommended Nodes (1/1)

Number of Clusters 1 HXAF-SP-220M5SNC-E [NVME][4TB] [Cluster 1]

2xIntel Xeon Gold 5220R Processor, 24 cores, 2.2 GHz | 768 [24x32] GiB DDR4 RAM (8 Base Sticks, 16 Add-On) | 3
Number of Nodes 2 + 1 (FT)
6x4.0TB, 2.5" NVMe | 1x375GB Optane | 1 RU

Number of RU 3

© 2020 Cisco and/or its affiliates. All rights reserved.

 Cluster 1 Overall Sizing Summary
Expected cluster utilization based on total resources required by all workloads and total available resources on this cluster
CPU RAM

Total Usable* CPUs = 430 vCPUs Total Usable* RAM = 1,392 GiB
(After System overhead, CPU overprovisioning ratio = 4, (After System overhead, RAM overprovisioning ratio = 1,
Performance headroom = Sustain 1 node failure) Performance headroom = Sustain 1 node failure)

Workload Unused Workload Unused

PU 240 vCPUs 189 vCPUs RAM 960 GiB 432 GiB

0 107
Best practice* sizing threshold: 387 vCPUs 215 322 429 0 348 696
Best practice* sizing threshold: 1,280.6 GiB1044 1392

Storage Capacity Storage IOPS

Total Usable* Capacity = 22.1 TB (20.1 TiB) Total available IOPS = 104K
Total Effective* Capacity = 30.7 TB (27.9 TiB) Workload Model: 8K IO Block Size, 70/30 Read/Write
(After Replication Factor = 3, System overhead,
Effective storage savings= 38.9%)
Workload Free Space Available Workload Unused / Performance Headroom
city 14 TB 16 TB Storage IOPS 6% 94 %

0 8 15 23 30
Best practice* sizing threshold: 21.5 TB (19.5 TiB) of Effective Capacity 0 25 50 75 100

© 2020 Cisco and/or its affiliates. All rights reserved.

* Definitions for Usable, Effective and Best Practice defined in the Glossary in the Appendix.
Fixed Config Sizing Report-cluster_1

© 2018 Cisco and/or its affiliates. All rights reserved.

User Inputs

Node Properties

Node Type CTO CPU Type (Cores, GHz) 5220R (24, 2.2)

HyperFlex Node Type HXAF-220M5SX RAM per Node (GiB) 768 [12 * 64GiB]

No. of HyperFlex Nodes 3 No. of Disk Drives per Node 6

Compute Node Type HX-B200-M5-U Disk Drive Size (GB) 3840

No. of Compute Nodes 0 Cache Size 1600 [NVMe]

Cluster Properties

Threshold Standard Replication Factor RF3

Hypervisor ESXi Performance Headroom N+1

Hardware Acceleration Off Compression Savings(%) 20

HyperFlex Boost Off Dedupe Savings(%) 10

© 2020 Cisco and/or its affiliates. All rights reserved.

Sizing Calculator Report Results

80 1,324.8 GiB 40K

Total Available Physical Cores Total Available Memory Total IOPS
[equivalent to 101
Intel Platinum 8164 cores]

21.2 TB (19.3 TiB) 29.4 TB (26.8 TiB) 20.6 TB (18.7 TiB)

Total Usable Capacity Total Effective Capacity Best Threshold Practice
[21.2 TB after 20% compression becomes
21.2 / 0.8 = 26.5 TB
26.5 TB after 10% dedupe becomes
26.5 / 0.9 = 29.4 TB]

© 2020 Cisco and/or its affiliates. All rights reserved.

Node Calculation
Compute (Cores) Memory (GiB) Disk (TB/TiB) Notes
# (HX + # (HX +
# HX Nodes * #
Compute) Nodes Compute) Nodes
144 2,304 Disk Slots * Disk 69.1 TB (62.9 TiB)
* # Sockets * # of * # DIMMs *
Size
Cores DIMM Size

Hx Sizer uses an Intel Platinum 8164 CPU as a baseline.

Factor for 5220R
CPU Normalization 178.9 Selected CPUs are normalized using the CPU SPEC ratings
[Cascade]: 1.24
against the baseline CPU

Total Physical
178.9 2,304 69.1 TB (62.9 TiB) Total physical resources available for sizing purposes
Resources

These are resourced used to run the HXDP stack that enables
Controller VM &
4 pCPUs / node -12.0 48 -144 8% -5.5 TB (-5.0 TiB) data services such as dedupe, compression, instantaneous
Metadata Reserves
snapshots, clones, etc.

Overprovisioning includes CPU and RAM Overprovisioning

Overprovisioning/ Ratios.
1 166.9 1 2,160 RF3 21.2 TB (19.3 TiB)
Replication Factor RF3 ensures that data is available even after the cluster has 2
uncorrelated and simultaneous failures

Performance N+X: This ensures there is enough performance capability to

N+1 -55.6 N+1 -720
headroom (N+X) sustain applications after X failures

Total resources usable after accounting for controller VM

Total Usable 112 1,440 21.2 TB (19.3 TiB)
overhead, and performance headroom

Best Practice CPU & Memory: This ensures predictable behavior for user VMs
10.0% -11.1 8.0% -115.2 30.0% -6.4 TB (-5.8 TiB)
Threshold Storage: Allows for better resilience and performance.

Storage Efficiency This multiplier shows how much storage has increased due to
1.39 20.6 TB (18.7 TiB)
Savings Multiplier Dedupe and Compression Savings

Best Practice
101 1,324.8 20.6 TB (18.7 TiB) Includes Best Practice Not-to-Exceed Threshold
Resources

© 2020 Cisco and/or its affiliates. All rights reserved.

Assignment of Workload to Clusters (1/1)

Number of Clusters Created: 1

Workload Type Description Count RF Fault Tolerance Cluster Assignment

VSI-1 VSI Medium 60 3 N+1 cluster_1

© 2020 Cisco and/or its affiliates. All rights reserved.

Overall Sizing Results

Sizing Summary Recommended Nodes (1/1)

Number of Clusters 1 HXAF-220M5SX [Cluster 1]

2xIntel Xeon Gold 5220R Processor, 24 cores, 2.2 GHz | 768 [12x64] GiB DDR4 RAM | 6x3.84TB, 2.5" SSD | 1x1.6TB 3
Number of Nodes 2 + 1 (FT)
NVMe | 1 RU

Number of RU 3

© 2020 Cisco and/or its affiliates. All rights reserved.

 cluster_1 Overall Sizing Summary
Expected cluster utilization based on total resources required by all workloads and total available resources on this cluster
CPU RAM

Total Usable* CPUs = 446 vCPUs Total Usable* RAM = 1,440 GiB
(After System overhead, CPU overprovisioning ratio = 4, (After System overhead, RAM overprovisioning ratio = 1,
Performance headroom = Sustain 1 node failure) Performance headroom = Sustain 1 node failure)

Workload Unused Workload Unused

PU 240 vCPUs 205 vCPUs RAM 960 GiB 480 GiB

0 111
Best practice* sizing threshold: 401 vCPUs 223 334 445 0 360 720
Best practice* sizing threshold: 1,324.8 GiB1080 1440

Storage Capacity Storage IOPS

Total Usable* Capacity = 21.2 TB (19.3 TiB) Total available IOPS = 61K
Total Effective* Capacity = 29.4 TB (26.8 TiB) Workload Model: 8K IO Block Size, 70/30 Read/Write
(After Replication Factor = 3, System overhead,
Effective storage savings= 38.9%)
Workload Free Space Available Workload Unused / Performance Headroom
city 14 TB 15 TB Storage IOPS 10 % 90 %

0 7 15 22 29
Best practice* sizing threshold: 20.6 TB (18.7 TiB) of Effective Capacity 0 25 50 75 100

© 2020 Cisco and/or its affiliates. All rights reserved.

* Definitions for Usable, Effective and Best Practice defined in the Glossary in the Appendix.
Backup

© 2018 Cisco and/or its affiliates. All rights reserved.

Workload Summary (1/1)
Workload Type Details Cluster Parameters

VSI-1 VSI Profile Type- Medium | Number Of VMs- 60 | vCPUs per VM- 4 | RAM per VM(GiB)- 16 | Average IOPS Replication Factor- 3
per VM- 100 | User Data Capacity per VM(GB)- 200 | Base Image Size(GB)- 20 | Number of Snapshots- Fault Tolerance Nodes- 1
5 | Compression Savings (%)- 20 | Dedupe Savings (%)- 10 | Working Set(%)- 10 | vCPU
Overprovisioning Ratio- 4 | Cluster Type- normal

© 2020 Cisco and/or its affiliates. All rights reserved.

Thank you.