When many people think of the cloud, they think of fast, low-cost deployments of relatively simple applications, not complex, game-changing systems like supply chain management. Supply Chain Nation asked Serge Massicotte, JDA’s chief technology officer, for his thoughts on this matter. In our first post on this topic, Massicotte discussed the many benefits of cloud deployment. In this post he will debunk some myths about SaaS applications and provide some suggestions on cloud-based deployments.
SCN: There has been a perception that cloud deployments, and especially SaaS applications, are not suitable for large, complex environments and systems such as a WMS. Has that changed?
Massicotte: Yes, I think it has changed significantly over the past seven years. That perception comes from an era when SaaS applications were addressing the SMB market. In that case, SaaS companies had to have multi-tenant systems because they needed to have thousands of customers sharing the same servers for economy of scale. Going back to 2007, I was the general manager of the platform at Taleo, a talent management system offered as a SaaS application. Taleo was going after large enterprises. Analysts agreed that SaaS was a good business model for SMB, but they were not as sure for large enterprises. I think things have completely changed since then. When I left Taleo in 2012, 70 of the Fortune 100 companies were customers and we had several hundreds of customers from the Fortune 5000. Taleo proved that SaaS could work for large enterprises— in a profitable manner –where environments are very complex. There are many other examples of SaaS companies that have addressed the needs of large customers. Let’s not forget that JDA has over 200 customers on our cloud solution and it is our fastest growing business segment. So I think the market is ready.
One thing I have to mention is the concern about data security. SaaS is more robust than ever. There are best practices for developing SaaS applications and you have many companies providing services to do penetration tests on applications—basically security testing—and the quality of SaaS systems are very adequate. With that said, security should always be a critical element of consideration when selecting a software provider.
From JDA’s point of view, high availability and performance are paramount – especially for execution applications such as TMS and WMS. Data centers from tier one providers are very reliable, with up-time of 99.95% or above, which is probably better than many internal IT systems. The major players like Amazon, IBM and Microsoft are investing billions of dollars in their data centers around the world. With all of that investment, it is really difficult – if not impossible – for corporate IT to match an equivalent level of capability. I think we are at a point in time where all companies should leverage, to some extent, that investment by the major providers. At the end of the day, companies can put more focus on cost-benefit as CIOs are under unprecedented pressure to drive down costs relative to IT.
SCN: What recommendations do you have for companies considering moving their retail or supply chain applications into the cloud?
Massicotte: Like any other project, it is important to manage technical as well as non-technical items. Let me start with the technical items. For instance, one of the first steps is the implementation of the software in our environment. That has to be planned judiciously such that we are making sure to have an equivalent system, but optimized for our cloud. For example, that means taking care of configuration and integration with third-party systems, and additional tuning of our software during configuration. In addition, it includes considerations for moving data between our cloud and their system, and vice-versa, due to integration between systems. And that requires some IT work on their side that must be managed.
Another consideration, from a data perspective, is moving data from a customer’s premise to our cloud. This step is done one time at the beginning. Typically it doesn’t take that much time to move the data from their system into our cloud. But if the data batch is very large, we may plan a few days to do this. The biggest part of this task is often the customer setting up the batches and interface for the data to move from their premises into our cloud. You are going to see some latency and they need to assign IT resources on their side. Depending on the application, we may also move data between customer systems and our environment on a daily basis, but this typically does not cause latency problems and provides benefits by having the right data in the right place at the right time.
I think another point that is important is that the customer needs to understand the distribution of their users around the world and the ways they access the system. We have seen customers who have all of their customer traffic routed through a single exit point within their network, which can have a real impact on performance. So the nature in which they serve their end user and how they have their network set up can have an impact on performance that has nothing to do with our cloud environment.
Another consideration is how they want to interact with us. Are they going to provide the first line of support internally? Do they want us to do the first line of support? How do we communicate between their internal IT and support staffs with our support line? From a process standpoint, this needs to be discussed between JDA and the customer.
One last point—change management is crucial when you change the way people are doing things. For example, IT may not be used to deploying a cloud solution. It’s important to make sure our customers are aware of what needs to happen and to ensure their IT staff, and any other people who manage the system, are comfortable with our cloud solutions.
SCN: Thanks for your insights, Serge. In our final post on cloud deployment launching August 12, Massicotte will provide his vision of what all of this might mean for the future.