Updated on 2023-10-30 GMT+08:00

Advantages

With service development, an increasing number of enterprises choose to combine IoT technologies for business growth. Huawei Cloud IoT services have outstanding advantages in capabilities, costs, O&M, security, and ecosystem compared with MQTT clusters managed by enterprises.

Table 1 Comparison

Dimension

Item

Huawei Cloud IoT

MQTT Cluster Managed by Enterprises

Capabilities

Flexible protocols

Supports mainstream IoT protocols and proprietary protocols to meet requirements of different devices and access scenarios.

Provides the plug-in mechanism to parse custom protocols.

Supports only the MQTT protocol. The capability of supporting other protocols requires development. It is difficult, expensive, and inefficient to maintain multiple protocols.

Access

Provides series of multi-language, open-source IoT device SDKs.

SDKs are pre-integrated in popular modules and chips for multi-network and multi-protocol access. This simplifies device access and shortens the access time to hours.

Developers are required to be familiar with different programming languages, causing heavy development workload.

Performance stability

Supports smooth and elastic expansion of service resources after purchase.

Supports secure and stable connections of hundreds of millions of devices, reliable communications with 100,000 TPS concurrency, and devices going online concurrently with tens of thousands of TPS.

Ensures 99.95% service availability.

R&D engineers need to perform tuning. To ensure 99.9% or higher availability, R&D engineers who are proficient in open source MQTT and senior architecture personnel are required.

Features

  • Cell-based technologies control the fault scope.
  • Message tracing facilitates fault locating and cause analysis.
  • Supports device shadows.
  • Supports over-the-air (OTA) upgrades.
  • Supports product models, which abstract and summarize product functions to decouple software and hardware development and improve system integration efficiency.
  • Supports the plug-in mechanism to parse custom protocols.
  • Supports data forwarding rules. Data can be forwarded to more than 10 cloud services.
  • Supports device linkage rules. Rules can be customized based on time, conditions, actions to configure scenario linkage and implement automatic collaboration across applications, subsystems, and devices.
  • The open architecture is used to leverage cutting-edge technologies and services of cloud computing.
  • Extensive functions and solutions have served many customers in different industries.

Open-source MQTT provides basic functions. Developers need to develop a complete solution based on open-source capabilities. However, some open-source code left unmodified during intrusive modification by developers may cause accidents on the live network during open-source middleware upgrades.

-

-

-

Technical support

Provides 24/7 professional support.

The service ticket system responds within 10 minutes.

Open source MQTT does not provide technical support and has a large number of default configuration parameters. Enterprises need to adjust the parameters based on service scenarios. If developers of enterprises are not familiar with the open source code, their improper parameter adjustments pose huge potential risks to commercial systems. When problems occur, they have to solve problems by themselves.

Costs

Server cost

Servers do not need to be purchased.

Servers need to be purchased.

Labor cost

No extra labor is required for cloud services.

Enterprises need to pay for professional development and O&M teams.

Resource use

Resources are out-of-the-box and elastic for service growth and scale-out without interruption.

Enterprises need to develop the elastic resource scaling function by themselves.

Architecture cost

The high-availability, high-performance, and secure architecture is built based on cloud native 2.0 and supports continuous evolution.

It is difficult for an enterprise to build the architecture that achieves high availability, high performance, and high security.

O&M

Infrastructure O&M

Provides unified O&M, quick response, scaling, upgrade, and troubleshooting based on professional teams.

Enterprises need to build their own O&M teams or use third-party O&M teams to solve scaling, upgrade, and O&M problems. Statistics show that most service faults are triggered by scaling and upgrade operations. The O&M cost is several times or even dozens of times the development cost.

Service platform version

Provides unified update by public cloud service providers, fast version iteration.

It is managed by enterprises.

O&M

  • Provides full-link log analysis and message tracing.
  • Provides real-time monitoring and sensing of device statuses.
  • Supports custom service metric alarms.

It is managed by enterprises.

Security

System security

Establishes a trusted security system: It obtains international security certifications such as ISO27001, ISO27017, ISO27018, and CSA STAR. It complies with EU's GDPR.

  • Transport network layer: Border security protection is provided based on web application firewall (WAF) and distributed denial of service (DDoS). Efficient, secure transmission protocols such as DTLS, TLS, HTTPS, CoAPS, and MQTTS are provided.
  • Device side: A unique digital certificate is provisioned for each device for secure access. LiteOS-based security capabilities are provided.
  • Platform side: Threat analysis is performed on entire Huawei Cloud network. Huawei Cloud security service products and public security services or components are fully reused to build a security defense system.

It is managed by enterprises. End-to-end security is a systematic project with high requirements. It is expensive and difficult to build and maintain system-level security capabilities.

Data security

Provides a complete security protection system. Data is stored in redundant mode in the data center of the cloud service provider, ensuing data security.

Enterprises need to build data redundancy, backup, and recovery capabilities.

Disaster recovery

Supports active-active service deployment, multi-data center DR, as well as high availability and DR capabilities based on multiple regions and availability zones (AZs).

Self-managed clusters usually do not have DR capabilities. Huge investment in active-active service deployment and DR devices often ends up with low return on investment (ROI).

Vulnerability fixing

Establishes a comprehensive vulnerability management system and a dedicated security research department to detect, track, and fix vulnerabilities in a timely manner.

Most enterprises do not have a vulnerability management mechanism or fix vulnerabilities in a timely manner. As a result, they are prone to attacks and are not aware of attacks and data theft.

Ecosystem

Third-party access

Integrates upstream and downstream ecosystem resources and provides value-added services.

It is built by vendors.

Scalability

1. The platform supports fast scale-out of tens of thousands of devices to hundreds of millions of devices without service interruption.

2. When other functions, such as AI, are required for service development, the platform can be seamlessly interconnected with other Huawei Cloud big data, EI, and middleware products to implement storage, computing, and intelligent analysis of device data at scale. In addition, cloud-based products support small-scale verification, facilitating fast, low-cost trial and error and service innovation.

The scaling period is long. Enterprises need to implement system or component interconnection. The labor and equipment costs are high.

Table 2 Expense comparison

Item

Huawei Cloud IoT

MQTT Cluster Managed by Enterprises

Cloud resource cost

An SU1 allows up to 400,000 messages per day and costs USD25 per month or USD300 per year.

Total: about USD300/year

  • Server resources: Two ECSs (AP-Singapore region, x86 architecture, general computing, 4 vCPUs, 8 GB memory, and 40 GB high I/O disk) cost USD2102.64 per year.
  • Relational Database Service (RDS): A general-purpose DB instance (MySQL engine, 2 vCPUs, 4 GB memory, 40 GB cloud SSD disk, and primary/standby type) costs USD890.56 per year.
  • Elastic Load Balance (ELB): An instance (AP-Singapore region, pay-per-use, public network, shared load balancer, and 1 Mbit/s bandwidth) costs USD262.8 per year.

Total: USD3256/year

Labor cost

None

Basic middleware implements basic functions.

  • One engineer is required for routine O&M and R&D of the platform.
  • Assume that the engineer devotes 50% of efforts and the monthly salary is USD10,000.
  • Total: CNY10,000 x 12 x 50% = USD60,000/year

Special functions are added based on the basic middleware.

  • Assume that only some functions are implemented without considering high availability, high performance, and high security of the platform.
  • Two full-stack development engineers are required for the platform frontend and backend development and O&M of functions such as device management, message communications, and rules.
  • One protocol professionals are required for device-side development, including implementing device access through native protocols, generic protocols industry protocols, and SDKs.
  • Assume that all engineers devote 100% of efforts and the monthly salary is USD10,000.
  • Total: 3 x CNY10,000 x 12 x 100% = USD360,000/month

Bonuses are not included.

Total

USD300/year

Basic functions: USD63,256/year

Basic and special functions: USD423,256/year