The requirements of business networks have changed mightily since the dot com era. Back then, company networks were primarily concerned with data communications and workstation connectivity. Today, businesses make more demands on network bandwidth than ever before.
Fifteen years ago, most businesses relied on the public switched telephone network (PSTN) to make calls. Now, much of this traffic has migrated to Voice over Internet Protocol (VoIP) applications. More recently, IoT has irrevocably transformed network communications as businesses integrate IoT devices into their networks.
As we know, IoT devices come in all shapes and sizes. From cameras to radio frequency identification (RFID) chips to digital signage, IoT devices bolster productivity and improve communications. But network designers and administrators find themselves scrambling to accommodate these devices on networks not built for IoT applications. Look no further than IPv4 exhaustion.
IoT endpoints can be highly interactive (viz., smart security systems) or completely passive (viz., RFID chips). Your company’s networks must accommodate not only host both IoT and NoT (non-IoT) devices but various IoT technologies including Bluetooth, Zigbee, and Near-field Communication (NFC).
Likely, your company utilizes either software-defined wide area network (SD-WAN) or Multiprotocol Label Switching (MPLS) for business communications. How do these architectures benefit IoT applications? Let’s dive into the debate of SD-WAN vs MPLS for IoT applications.
MPLS Versus SD-WAN for IoT
To be clear, SD-WAN is a software-delivered strategy providing multi-cloud connectivity to users. MPLS, on the other hand, utilizes routing tables and label switch routers (LSR). MPLS is a well-established technology, introduced in the 1990s. SD-WAN is a relatively new approach, rapidly superseding MPLS because of its (typically) lower cost and intrinsic flexibility.
Cisco originated MPLS, handing off the technology to the Internet Engineering Task Force (IETF) for open-source standardization. MPLS is a hardware-based solution, dependent on very large-scale integration (VLSI) circuits. SD-WAN, as its name implies, uses software to route traffic over Internet broadband. SD-WAN is highly adaptable to a multi-cloud environment; MPLS is not.
SD-WAN has been described as “a backbone for IoT deployments.” Conversely, MPLS — widely deployed by telcos — is “ill-equipped” to serve the breadth of endpoints over an IoT system. Let’s see why network architects prefer SD-WAN over MPLS for IoT.
IoT Considerations When Comparing MPLS and SD-WAN
While carrier-based MPLS is a tried-and-true technology with an outstanding quality of service (QoS), interconnectivity is limited to a private dedicated network. Once a business expands beyond the MPLS footprint of the carrier, costs, scalability, and deployment time become untenable. Imagine the cost of connecting to every IoT device in your network using private MPLS circuits.
SD-WAN, as mentioned, connects to IoT endpoints using the public Internet. Internet broadband is much more cost-effective than carrier VLSI circuits. Moreover, SD-WAN is a cloud-based solution addressing software as a service (SaaS) connectivity. MPLS does not work in a multi-cloud environment without backhauling, impacting QoS considerations like latency, jitter, and packet loss.
Happily, for enterprises deeply invested in MPLS, many SD-WAN solutions allow for connectivity over MPLS while diverting less sensitive traffic to the Internet. Thus, businesses can maintain service level agreement (SLA) guarantees for prioritized data while utilizing SD-WAN for non-time reliant or broadband (viz., 25 Mbps upstream/3 Mbps down) IoT applications.
When assessing a network’s suitability for IoT devices, network architects evaluate the following:
- How many IoT endpoints will be deployed?
- Is the endpoint mobile or fixed?
- Does the network control activity at the endpoint or is the device passive?
- What security protocols does the endpoint require?
- Does the device generate data requiring real-time analysis?
- Is the device battery-powered or does it require external power (viz., PoE; Power over Ethernet)?
- What are the IoT network reliability requirements?
- Do IoT devices need application customization or configuration?
- What are endpoint data and traffic management and monitoring requirements?
- How much does large-scale IoT connectivity cost?
While some IoT devices have specific upstream connectivity needs (viz., video surveillance), many IoT use cases require only infrequent connectivity with relatively small bandwidth. In such non-mission-critical cases, deployment of Low Power Wide Area Networks (LPWAN) or NarrowBand IoT (NB-IoT) is appropriate.
In some applications, standardized cellular technologies can be used for connectivity. However, past mobile solutions relied on obsolete 3G network protocols. As we know, mobile network operators (MNOs) are rapidly decommissioning their 3G networks to make room for LTE and 5G.
Thus, future mobile interoperability using 3G standards will soon end. For networks requiring mobile capabilities and low data throughput, options include long-range WAN (LoRaWAN), DASH7, and the proprietary Sigfox. But these technologies do not address the connectivity needs of data-intensive computing applications generating datasets from multiple locations and in various forms.
So, to state the obvious, a company’s unique needs — reliability, latency, security, data management — dictate how IoT is administered over business networks.
Is SD-WAN or MPLS Better for IoT?
Undoubtedly, SD-WAN is the choice for most businesses integrating IoT endpoints into their company’s networks. A crucial consideration for choosing SD-WAN for IoT is that it allows network administrators to centralize access, enabling them to manage all connected endpoints — including BYOD devices — from one location.
For those wondering about the network security of SD-WANs and the cloud, new technologies have emerged to address this concern. Next-generation firewalls (NGFWs) and secure access service edge (SASE) architecture allow network administrators granular visibility to SD-WANs. This decentralized approach to network security uses software upgrades, not new hardware, to lower cybersecurity costs.
Those interested in an SD-WAN solution for their business needs can choose from multiple vendors. They should make their selection(s) based on their IoT application requirements and performance criteria (viz., reliability, network capacity, data rates, payload sizes, etc.) An SD-WAN solution avoids both vendor- and carrier- lock-ins, making it a versatile option for IoT.