Continuing on my thoughts around taking an application-centric view and the benefits of adopting 5G, one such application domain that we should consider is Enterprise Resource Planning(ERP). ERP is probably the only business management application platform that governs every aspect of any enterprise operation. While the telco industry is touting the benefits of 5G for the enterprise, let us go a little deeper into the various functions within the enterprise which are handled by an ERP system and really understand if 5G will meet the needs or whether if it is even required.
The picture below presents the various modules that are part of an ERP platform (reference: Oracle ERP solutions).
Now, by no means am I the expert in the ERP domain(this disclaimer helps me not to write about it), and I would welcome a debate around the context that I am about to set in this blog.
5G benefits for an Enterprise
At a high level, the 5G benefits for any enterprise are as follows (source: IMT -2020 ITU) →
- Support for high bandwidth — with the flexibility to choose your multi-gigabit throughput, the peak data rate of 20Gbits/second.
- User Capacity — 5G will support a larger user density compared to its predecessors (LTE) and WIFI.
- Managed latency — The flexibility to define what latency and QoS we need at a use case level, based on flexible subcarrier spacing and use of different spectrum bands — FR1 (low and mid-band) and FR2 (millimeter wave), with an ultra-low latency target of 1ms.
- More reliability and enhanced security → reliable and uniform user experience. With a User experience data rate of 100 Mbits/sec
- Increased availability — indoor coverage and outdoor coverage roaming, with mobility support of 500 km/h, and traffic capacity handling of 10 Mbits/sec/sqmt.
- Flexible deployment options for enterprise with separation of Control plane and User plane. Where user plane (data traffic) can be deployed at various enterprise sub-locations all controlled by a central core.
Now, of what use is this to an ERP implementation?
Any enterprise, be it small medium, or large global enterprise will have use-cases that will be classified as below :
- High bandwidth, high-speed connectivity — most of the IT application data, personnel access data would fall under this category. We classify this category as eMBB — or enhance mobile broadband.
- Low bandwidth, frequent data — these are small sensor devices that are used to track, or monitor assets, and asset performance, etc. Small amounts of data transmitted frequently. We classify this category as MMTC — or machine to machine transmission communication, finally
- Extreme low latency, with the flexible volume of data, high response, and control — this is applicable for that system that depends upon extremely fast response time measured at millisecond levels. Particularly applicable in industrial segments like manufacturing or utilities like power grids etc, this is classified as URLLC — Ultra-reliable low latency.
The picture below shows the use cases within the ERP domain that may fall under the above three categories — eMBB, mmTC and/or URLLC. Chances are each use case may fall under various combinations.
Each of these three categories has different 5G Quality of Service (5QI) values. 5G architectures and the standards define the guidelines for these 5QI values, and the cellular carrier may configure those depending upon the enterprise use case.
For example, in the picture above, I have identified, uRLLC, eMBB, and mmTC to be applicable for manufacturing management. This is purely my opinion and happy to debate.
Example use case within a manufacturing function
Individual use cases and devices within the manufacturing segment like — Programmable control systems (PLCs) may use URLLC, versus other machine traffic which includes sensor data may use eMBB or mmTC depending upon how critical is the use case.
In the case of PLCs which are control devices, they will have a higher dependency on low latency, and reliability of that connection. Therefore all PLCs may have to be segmented to use mmWave Radio setup to achieve near 1 ms latency, which will not be available at a sub 6 spectrum range (Read this to understand how latency is achieved in 5G).
The segmentation of this traffic will be done using 5G Slicing. With a dedicated slice allocated for that manufacturing plant, the PLCs and the traffic will be mapped to a specific SDF ( Service data flow) and will be assigned a specific 5QI ( 5G quality of service Indicator) → (See this !)
This same strategy may continue to be applied for other use cases falling within the ERP domain.
Below is a table describing, in general, the network requirements for use cases utilizing each ERP feature, and where 5G will fit in.
In the picture above the majority of the use cases for low latency fall into the manufacturing bucket (marked in red)
About WIFI 6
So now you may be wondering, why shouldn't we use WIFI 6. You absolutely can. WIFI is one of the incumbents and will be relevant forever. WIFI 6 is the new standard utilizing the same technology (OFDM) as 5G and offers some of the benefits. But the main obvious benefit of using WIFI 6 is that the spectrum is unlicensed.
If you look at the table above, the majority of the use cases within ERP fall under unlicensed.
Until now Industrial wifi has co-existed very well with fiber implementation. But now, with the maturity in cellular networks especially with 5G, the battle for the enterprise has just begun. The debate will be around what does the enterprise want.
Some key decision points while evaluating WIFI vs 5G would be —
1.Support for Global mobility
— Enterprises with worldwide operations benefit if one cohesive network fabric infrastructure drives their entire operation. Within this cohesive network, each function within the enterprise is divided/segmented. This is existing today in some form, especially with fixed corporate networks with most indoor functions on Wifi or fixed networks, and outdoor functions on Cellular. This comes with high operational costs of maintaining and managing three different networks. Cellular clearly has a chance to win here.
2. Cost of Indoor coverage
Cellular obviously offers higher coverage area and also supports a larger device density per square foot, compared to WIFI 6. Very specifically for Enterprise wifi, in an industrial setting covering 300,000 sq ft of floor space, and 2500 plus device, per sq foot cost, is under $50, with full managed operations and an SLA of over 99.99 %. Cellular will have to beat that. This will take some time, and some heavy selling into the enterprise.
3. Latency dependency
This will depend upon the use case. From the table above, within the ERP domain, the only use cases which really depend upon URLLC would be in an industrial setting. Depending upon the use case the latency tolerance will vary. FR2 (mmWave) spectrum will play a larger role here. And currently almost all of that spectrum today is licensed. Therefore the carriers will have dominance here.
4. Advancements in Network slicing
This is a mature and very compelling value proposition with 5G. The thought leadership around this specific topic is evolving and a lot of progress is being made to orchestrate end-to-end slicing. If there is a seamless orchestration enabled by the carriers for an enterprise to cover all types of use cases and allocating 5QI defined service data flows within those slices without the Enterprise ever have to worry about dimensioning, configuring, and maintenance of their networks, we have an obvious winner over Wifi or Wifi 6.
I think we have to really ask the question “what does the enterprise want from 5G”. The industry has done a fine job touting 5G all over the map and the Enterprise has woken up to understand the benefits 5G claims to deliver. But unless and until the industry takes an application-centric view and “integrates” the value 5G offers within its application space, eliminating the enterprise to view 5G as just another network bearer we will see sluggish improvements in its adoption. Especially in the brownfield implementations.
I am specifically interested in helping the enterprise should not think of the network as its own siloed existence. All I want to see is that the Enterprise buys an application platform and the network is implicit. Whether it is 5G or Wifi or fixed ….it’s just an option that the enterprise selects. Rest everything is programmable to support different use cases.