Running core systems on-premise was not a feasible long-term plan for Coca-Cola Amatil, an organisation that manages huge warehouses and hundreds of thousands of transactions every day.
Speaking at an IBM event in Sydney today, CIO Barry Simpson gave a run down of the Coca-Cola bottler's transition to the cloud.
CCA recently signed a five-year multimillion dollar cloud deal with IBM to move its core systems to the cloud and streamline its order management and distribution processes across the South Pacific region. The Coca-Cola bottler is moving its p-Series platform and around 100 LPARs (logical partitions) to the cloud.
IBM’s data centre in Sydney will act as a cloud hub for the rest of CCA's South Pacific operations.
Simpson said his focus this year is on growing CCA's market share in Indonesia, which the cloud will play a key role in being able to support large volumes of equipment and transaction data out in the field.
CCA manages more than 300,000 pieces of equipment, such as fridges and kiosks. “We try to do that better than our competitors, so rock solid transaction processing is extremely important to us,” said Simpson.
“We regularly achieve over 98 per cent of our orders delivered in full, on time and accurately invoiced. And that's with an SKU portfolio of over 1,500 SKUs [stock keeping units], over 150,000 distribution points of orders taken 24 hours a day.”
Simpson expects to see significant savings by moving to a consumption-based cost model, where the cloud provides flexibility to ramp up computing power only when it's really needed.
“As we are extending our platform to our Indonesian business where we are servicing 500,000 customers... you can't afford to get those businesses onto that platform unless you adjust your cost base, [make it] based on consumption,” he said.
Read more: Cloud Computing Survey
CCA is also looking to enhance its business intelligence capabilities using IBM SmartCloud. Simpson said he currently collects point of sale data on the 300,000 plus pieces of equipment that are geo-tagged, and also works with customers to provide them insights from this information. However, the internal reporting side of this needs improving, he said.
“Today when we look at the profitability and performance of that equipment, it's all done in static reports. And when you provide it to someone they say 'I have another question, and therefore I need another report', we say 'good news, we'll see you in six months and come back with the next report'.
“So we are moving to in-memory processing, to real time data, and we are delivering it directly to that mobile end consumer [salesperson].
"Instead of having people out in the field collecting data and feeding it into the [head management] centre, we are feeding it out to the people who can actually make use of it and doing it way faster than we could have traditionally done.”
About 1,000 salespeople, who work out in the field with iPads, will be able to access the datasets, do the analysis, and have it visualised to be able to easily interpret it and gain insights, Simpson said.
They can also share their insights through CCA's social Intranet.
“If you really believe that insight should lead you to the next question, then your approach to BI has to be very different. In my view, you can't win by just changing or trying to update the current game; you have to actually change the game,” Simpson said.
Simpson said he doesn't see any need to keep any of CCA's IT systems on-premise. He pointed out that organisations must be assured cloud providers offer adequate security, and they can still have control and visibility over their data. But that doesn't necessarily mean organisations need to be burdened with the task of maintaining all their systems in-house in order to be secure.
“To deliver those services and deliver innovation, we needed pace. My view is certainly all roads lead to cloud. I don't think you can get there by keeping your systems on-premise. In the near future, I don't see a need for having any of our systems on premise.”
- Govt should adopt ‘cloud first’ policy
- GPT's breakneck cloud transition
- Feature: Why mainframes are still around