By Sramana Mitra and guest author Shaloo Shalini
SM: Take me through how a SaaS vendor like MercuryGate sits in this. Is MercuryGate receiving data through Liaison, or do you send data to MercuryGate?
PS: Very good question and a great use case. We do both: We send them data and they send us data. Let me take you through this. Our enterprise resource planning system, which is on-premise, sends loads to Liaison every fifteen minutes. We have a shipment that load information goes to Liaison, they do a transformation, and they [use file transfer protocol], or FTP, to send it in XML format to MercuryGate. On the way back every night, we get a file from MercuryGate through Liaison, and they just pass it through us, and that represents the invoicing. Every night. Meanwhile, MercuryGate is collecting all EDI transactions from these 30 or 40 transportation carriers. Unknown to us, they handle all that. So, there is a bilateral integration rolled out in real time, and an invoice file comes back to us. That is a public cloud to private cloud integration through Liaison.
SM: Very interesting. What is that cost of integration, what does that add to your cost structure?
PS: What is interesting about that is, again as I mentioned, our on-premise transportation management system failed because we could not handle integration on-premise. It was too complicated – it was Java based, and we could not handle that. In my view, using this combination of cloud-based integration as a service provider and a SaaS application that has deep domain-level expertise, it mitigates if not completely eliminates the integration problem. The cost of this integration was a couple of thousand dollars, maybe.
SM: Wow, and it is working really well?
PS: Sure enough. You probably know that 15 years ago when we took on these big projects, one-third was hardware and software costs, one-third was integration costs, and one-third was built-in cost. That model has shifted and gone, integration costs should go away, and infrastructure costs are going away. It’s all about licensing and software as a service, which is now a fixed cost.
SM: Very good.
PS: Let me emphasize that integration is the biggest single impediment to adoption of cloud computing – it is this idea of effectively doing integration, doing it for a mass and within a range of complexities.
SM: I presume that over the next few decades you are going to add a lot of technologies, a lot of other cloud technologies such as SaaS or IaaS to your portfolio, just because there is so much technology available at such an affordable cost at this point. So, you need a way to be able to bring that on board, integrate it, and manage the data transfer between systems.
PS: Exactly, and I have the freedom of choice: Should I do this on-premise or off-premise? That should not be a significant variable in that decision.
SM: What are your thoughts about ERP? you mentioned that ERP systems are on-premise. I hear that a lot of SMEs are moving their ERP system out to NetSuite and something like that in the cloud. Is that part of your plan, or is there some exit barrier there?
PS: Yes, that is part of plan, and our ERP vendor is Encore. We very much want to move it to the cloud. The problem is that they are not ready. They need to service-enable their application, they need to Web-enable their application in order to move it out as a SaaS solution. So, the if legacy was green screen report program generator (RPG)-based IBM iSeries, they are rewriting their software to be more modular, more service oriented, and more Web based. Then it flows naturally – this will become Web enabled and can be hosted.
SM: You are not looking at a provider that is ready and has a large footprint already?
PS: No, because there is so much institutional knowledge invested into ERP packages. Part of an interesting phenomenon is this idea of vendor lock-in. By this I mean monolithic enterprise systems. We did this decades ago, and now we are stuck with them. But we are slowly beginning to decouple it – the enterprise system here, the CRM there, and it is slowly starting to break apart. Over the next couple of years, the size and scope of our enterprise system will get smaller, and it will be easier for us to move parts of it to off-premise. But to answer your question, we could not afford to adopt a new ERP system. We have to stay with the one we have. But eventually, it will be a hosted external app.