From an ecosystem point of view, regardless of the protocol used at a sensor level, the sensor data will ultimately be fed into a public, private, or hybrid cloud for analysis, control, or monitoring. Outside of the WPAN, the world is TCP/IP-based, as we see in WLAN and WAN configurations.  

IP is the standard form of global communication for various reasons:

• Ubiquity: IP stacks are provided by nearly every operating system and every medium. IP communication protocols are capable of running on various WPAN systems, cellular, copper wire, fiber-optic, PCI Express, and satellite systems. IP specifies the exact format for all data communications and the rules used to communicate, acknowledge, and manage connectivity.
• Longevity: TCP was established in 1974, and the IPv4 standard still in use today was designed in 1978. It has withstood the test of time for 40 years. Longevity is paramount for many industrial and field IoT solutions that must support devices and systems for decades. Various other proprietary protocols have been designed by various manufacturers in those 40 years, such as AppleTalk, SNA, DECnet, and Novell IPX, but none have gained the market traction as well as IP.   
• Standards-based: TCP/IP is governed by the Internet Engineering Task Force (IETF). The IETF maintains a set of open standards focused on the internet protocol.
• Scalability: IP has demonstrated scale and adoption. IP networks have demonstrated massive scaling to billions of users and many more devices. IPv6 could provide a unique IP address to every atom comprising Earth and still support 100 more worlds.
• Reliability: IP at its heart is a reliable protocol for data transmission. It accomplishes this through a packet delivery system based on a connectionless network. The service is considered unreliable from conception, meaning the data is not guaranteed to be delivered. IP is connectionless because each packet is treated independently from one another. The IP is also referred to as best-effort delivery because all attempts will be made to transmit a packet through various routes. The strength of this model allows an architect to replace the delivery mechanism with another—essentially replacing layers one and two of the stack with something else (for example, Wi-Fi with cellular).
• Manageability: Various tools exist to manage IP networks and devices on an IP network. Modeling tools, network sniffers, diagnostic tools, and various appliances exist to assist in building, scaling, and maintaining networks.

The transport layer is also worth considering. While IP addresses the need for a well supported and robust network layer, TCP and Universal Datagram Protocol (UDP) are needed for the transport layer. The transport layer is responsible for end-to-end communication. The logical communication between different hosts and various network components is governed at this level. TCP is used for connection-oriented transmissions, whereas UDP is used for connectionless transmissions. UDP is naturally much simpler to implement than TCP, but not as resilient. Both services provide segment reordering as packets are not guaranteed to be delivered in order using an IP protocol. TCP also provides the layer of reliability to an unreliable IP network layer through the use of acknowledgment messages and retransmissions of lost messages. Additionally, TCP provides flow control using sliding windows and congestion avoidance algorithms. UDP provides a lightweight, high-speed method to broadcast data to various devices that may or may not be present or reliable. 

The following is the standard  7-layer Open Source Interconnection Model stack. 

From an IoT perspective, bringing IP close to the source of data bridges two worlds of data management. The Information Technology (IT) role manages the infrastructure, security, and provisioning of networks and things on the network. The Operational Technology (OT) role manages the health and throughput of the system that functions to produce something. These two roles have traditionally been separated, as things such as sensors, meters, and programmable controllers have not been connected, at least directly. Proprietary standards have governed the OT systems, at least from an industrial IoT perspective.