By Bill McBeath, Chief Research Officer, ChainLink Research
In my first blog post, I talked about three concerns that manufacturing firms have about moving to cloud-based solutions: security, availability and large enterprise support.
Let’s continue the conversation with a look at a few other common concerns, and the reality behind each of them.
Concern #4 – Performance (Latency)
Can a cloud MES keep up with the machinery on the shop floor?
Reality: There are hundreds of shops being successfully run directly on cloud-based manufacturing execution systems (MES).
There are two aspects to performance: scalability and latency. I’ll address scalability a little later in this post, but regarding latency, it is important to know the most stringent specific latency requirement that you have between the MES system and your shop floor equipment.
If the MES system is talking to a local integration server or SCADA system, you may not require a sub-second response time. But if the MES is interfacing directly to PLCs or shop floor equipment, latency requirements could be more strict — as low as 150ms or less.
For stringent latency requirements such as 150ms, geography matters. If the cloud provider’s data center is on the same continent as your factory and has good network connections, there should not be a problem. But if the provider is on a different continent — or if the factory is in a less-developed part of the country — a globally distributed dynamic caching system (such as Akamai’s Dynamic Site Accelerator) or similar mechanisms may be needed to ensure your latency requirements are met.
Assessing a Cloud Provider: Find out what QoS/latency guarantees are included in the offering and whether they match your needs. Ask where the provider’s data centers are and what type of global caching services they use. Include performance tests in your implementation and system acceptance plans.
Concern #5 – Functionality
Aren’t cloud systems really only suitable for light manufacturing and assembly? Will a cloud solution have the depth of functionality that we need?
Reality: It is true that many (but not all) of the cloud solutions in the market are intended for light manufacturing and assembly. Some of the areas in which they fall short for complex manufacturing include the depth of their data model, MES capabilities, machine integration and traceability (for more on this, see Manufacturing in the Cloud).
A notable exception is the Plex Manufacturing Cloud, whose functional footprint is quite mature, having been continually built out and deepened for more than 15 years. The Plex footprint includes:
- Scheduling – PRP, MRP, APS, Kanban
- Production – PLC machine integration, assembly line control, production and job tracking and management, workcenter management, OEE, setup, scrap, rejection, weigh scale
- Engineering modules – deep parts specification, engineering drawing/data management, process routings
- Substantial functionality in quality, traceability, tool management, maintenance, inventory, costing, procurement, shipping/distribution, EDI, HR, CRM, accounting and more
Assessing a Cloud Provider: List your most demanding requirements and check which of those they support. Ask for a demo to see how it works. Don’t forget to include future expected requirements as your business grows.
Concern #6 – Customization
Can cloud solutions be customized for our needs the way an on-premise solution can be?
Reality: This depends on the solution architecture. Look for a platform with forward-compatible configuration and customization. The architecture should guarantee that any customizations that are built according to the solution provider’s specific guidelines will continue to work on all future upgrades to the platform.
For example, Plex has a data-driven core that can be configured for all of the different manufacturing modes, processes and industries they serve. The data model allows for customization of workflows, screens, language, field names, lengths and other data attributes.
One major advantage of a cloud solution is that customers automatically gain access to new features as they are released, unlike an on-premise system where you have to pay for and go through an often-painful upgrade to get the latest release. With on-premise ERP, the system starts getting stale and dated; with cloud/SaaS, continual improvements keep on coming and you always have access to the latest advances.
Assessing a Cloud Provider: Pick one or more processes that you have some unique requirements for and ask the solution provider to show you how they will accommodate those requirements. Find out if and how they guarantee forward-compatibility with future releases.
Concern #7 – Scalability
Can a cloud solution scale up to support very large plants and large numbers of plants?
Reality: At a high level, the two scalability constraints to look at are the network bandwidth required and compute power/storage. The amount of network bandwidth required for a given workload is dependent on the architecture and design of the cloud solution.
In the case of the Plex Manufacturing Cloud, traffic per machine is fairly light and the network can be scaled up to serve virtually any size plant by increasing the throughput of the network service provided. For example: One of Plex’s customers is running a major plant with four lines and several hundred workstations.
Scalability of computational power is one of the key advantages of a SaaS approach compared to on-premise. With an on-premise solution, you need to buy enough hardware to handle your most intensive compute needs – but the rest of the time, that hardware is being underutilized. This means you are paying for horsepower that you are not using.
In contrast, with cloud ERP, the amount of compute power you need can be scaled up or down as needed, so you are only paying for what you are using — meaning cloud systems can scale to almost any size plant, provided a network service with the proper bandwidth has been selected.
Regarding scaling to a large number of plants: That is largely a factor of the design of the software, and how elegantly it allows the management and integrated view of a large number of plants.
Assessing a Cloud Provider: Ask the provider for an estimate of the network bandwidth needed to support your current plants as well as any future expansion you anticipate. This will enable you to price out the network service required. If you are a very large company with many plants, find out the current capacity of the solution provider’s data center and, if necessary, ask for a capacity assessment to make sure they can reliably scale up to meet your peak needs without impacting performance.
Bottom Line: Is a Cloud Solution Right for You?
Globally, manufacturing has become hypercompetitive, with the need to squeeze more profit out of every single piece built. Selecting the right technology is an important part of the answer to help a company grow and remain competitive.
If you are considering upgrading to an MES/ERP platform and you assumed that cloud systems could not support complex process-intensive manufacturing, you owe it to yourself to take another look. A cloud solution may be a good fit if you have limited capital, limited IT resources, and/or the need to flexibly scale up and down your infrastructure. Very capable cloud-based manufacturing systems exist now, and they have what is needed for today’s serious manufacturing demands and constraints.
Bill McBeath leads ChainLink’s research efforts, as well as the company’s procurement, strategic sourcing, design collaboration and online marketplaces practices. With more than 20 years of experience in a variety of roles as a business and technology researcher and consultant, high tech executive, and software architect, McBeath is recognized as a leading expert in extended-enterprise business models.