Delivery Architecture – What it Means…


Abstract needed here…



By now we hope you have been reading last issue’s articles Remapping the Supply Chain Universe and Mapping the End-to-End Supply Chain Process , as well as this issue’s Assessing Supply Chain Performance.

These will be prerequisites for understand the delivery architecture. In those articles we discussed the externalization of the business process—across multiple enterprises, and the challenge in managing your 3P’s (supply chain policy, performance, processes) using outdating techniques and systems.

Once we cross the border of the enterprise, traditional enterprise centric systems fail to provide visibility, or a way to understand and act, all in sync with our partners.

Delivery Architecture has many nuances (see On Demand report summary), but there are a few key issues to highlight. First, a large part of our intelligence today that is used in every day business decisions does not come from inside the enterprise (competitive data, market or supply data). So the billions of investments in ERP can only take you so far. To understand market dynamics, we engage a host of sources to create solutions that are a convergence of sensory data, internet subscriptions and searches, analytics and inference, transaction data, etc. And the person doing that is frequently not in a fixed location—i.e. amry personnel on the move, sales person visiting her customer, logistics person in a port. So, today’s systems are not just about the data, but also the infrastructure below, between, and beside the information (algorithms, transaction logic, etc.) that needs to be addressed in order to create today’s global solution.

Two Critical Architectures for the Enterprise

There are really two main information architectures that firms need to deal with.

Financial Network:

The fiduciary function maintains a system of record, both financial and operational, which aggregates and closes the books. This is the behind the firewall enterprise solution. Enterprise Cohesion is its goal. And these things are required by law.

Trading Network:

In addition, there is the real-time supply chain layer. This can be composed of both internal and external components—an ecommerce front end and supply chain planning internal enterprise component, and then the supply chain network. This total network requires real-time reasoning and the ability to seek out subscribers of information (publish/subscribe) to perform their roles based on process for the business interaction in play. Creating a concept such as this is fairly problematic with only behind the firewall approaches.

And it requires a policy change in sharing intelligence with your partners.

Financial Management Network

Networks are not just limited to supply chain, but can include any process in the value chain. The virtual model does extend to financial and other services. So there is a virtual financial model. Some businesses will still use ERP software for this. However, there are SmallSmartFast financial models in operation that will have broader reach as time goes on.

These models are enabled by the ubiquitous technology and have at their core agreements and standards. There are paradigms we are already familiar with. Think credit card and banking systems. The private individual today can be an eBay merchant—eBay operates the order management systems, the financial settlements, etc. You can even get your company benefits plan, health insurance plan, etc., by being an eBay merchant. Amazon has a similar program, not just for small business, but also for the likes of Nordstrom, etc.

In the supply chain arena there are organizations that can manage your financial settlements, own your inventory, forecast and deliver to your customer, and help with charge back and discrepancy settlements between retailer and supplier. We call these 3Pe business models (policy, process, performance, enablers). More and more firms are able to support cross channel operations, including the financial management elements of your firm. There may be an ERP system in the core to meet GAAP requirements, but also more and more businesses take advantage of high velocity cash, administrative outsourcing and the like. ERP as we know it today will undergo significant changes, if it wants to play in this arena. And this is not just about implementing SOA (Service Oriented Architecture) on top of your legacy systems. Reaching things with a browser does not, in of itself, make a system inter-enterprise.

Trading Network

Here is where the operations of the virtual firm are executed. My planning activities may be in my core—conducted for the benefit of revenue and strategic planning, but tactical supply chain planning is an element that I share with the network. S&OP can also be built on a cross-enterprise consensus platform—like real-time smart conferencing, simulation and reasoning.

So we need the functionality of the planning logic and a very fast interaction/information bus to conduct commerce—that is not being provided by the ERP systems we have today. Too many layers exist between reference data (customer lists, pricelists, etc.), so these have to be exposed to the trading network. This is where ONS and other data synchronization services come into play (read the article on understanding ONS). Some of the data can be managed at the network hub—aka network applications that are shared by trading partners. This architecture needs absolute scalability as well as speed to deal with global traffic of RFID, and instant messaging as well as rapid semantic remapping of the content of these transactions.

It also needs to manage the convergence of voice, video, radio frequent, Internet and wireless transactions.

The pervasiveness of this technology is critical, since it lays the foundation for global and mobile computing—and for all workers, especially in the kind of work that exists outside the building—logistics workers, sales professionals, claims adjusters and meter readers.

The Delivery Architecture

So, if much of today’s work is done outside the building, and beyond single enterprise, then we need an IT architecture that fits that. Today, commonly, we think about the implementation of Service Oriented Architecture—SOA.

But SOA is only part of the story. I can access ERP from a web browser, but that does not make the system inter-enterprise.

So buying technology to create the integration from the physical layer to the network can be confusing to the user.

In particular, most consideration should be given to the business processes, as well as the physical business topology that will operate the process, in order to think about whether you are going to buy traditional software licenses or buy your solutions stretched across the web.

Maybe a network application, rather than a license, is the way to go.

Stay with me here, because we are getting to why these systems will be so revolutionary in the supply chain or trading network over time. This way of looking at data source/input can have a profound effect on the kind of middle systems we desire for the trading network or even the financial network applications.

Beyond traditional thinking about middleware, the hub data model can be designed to have a single instant of the product record. As the product moves through the chain, the data object is appended to, adding status changes, etc. These triggers can then alert the ‘waiting’ destination systems, such as invoicing, etc. that the product has been successfully delivered and accepted—OK to invoice. Today we pass huge data packs around. Tomorrow SmallSmartFast signals with tiny bits of secured information can literally set a flag, and the whole network can be updated (publish/subscribe). This concept is important, because it documents information cycle time exponentially, and reduces traffic, which is a nice thing to think about, since there will be a huge increase of traffic with all these signals coming and going from all these devices. But most important—one secure signal can update many trading partners and set significant business activity in motion.

Quite frankly, this is not a new concept or technique. It is used everyday in the banking network. The bottom of your check has a few key numbers that tell all to the network. These networks are secure and n-scalable. However, the new network in supply chain is based on much newer technology approaches which allow some special capabilities—sensing—that the banking network does not allow today. (With the huge amount of credit card fraud, think about new techniques and applications in Sense and Respond in our security consciousness society.)

RFID data synchronization gets enabled as well from this concept, since it allows many firms to subscribe to a single valid source, a SVOT (single version of the truth), which is beyond the firewall, and frequently updated. Therefore, it is a most eliable approach, and certainly a best fit architecturally for true supply chain management.


Buying software is not just about functionality. Consideration needs to be given to how, where, and with whom we are sharing processes and data. Networks will not only service the concept of inter-enterprise connectivity, but a host of new models—teeny applications that are based on shared data—not transmitted data. Also, the network can be the platform for new federated business models, as we discussed in the last issue of Parallax View.

In the hyper RFID world, many people are missing the ONS or data synchronization or Network platform component. Yet we think this is where the real action will be. Wireless—we love it—but the mobile world does not work without the hub. When we are going behind the moon, we have to be able to contact with our mission control and get re-synched with the SVOT in order to be the supplier of provider of choice for today’s demand customers—all across the world.

Scroll to Top