So the question becomes: what makes an ERP system in the cloud to be truly in the public cloud? 

Vendors have their own definitions and understanding of ERP in the cloud that doesn't always match the CIO community's thinking. Having been a CIO for many years, here's my personal opinion about the five key attributes for a public cloud ERP.

In the legacy world, when we wanted to use any kind of enterprise software, we first bought the perpetual software license and the required hardware and then we had to install everything in our own data centers or in a facility provided by a third party. Today, buying a cloud-based SaaS solution requires that we simply sign up on a vendor's website. With a click of a button, we can access SaaS software as long as we have connectivity to the Internet. 

Users don't really care where the data centers are located as long as their requirements are met. While the loss of control created some initial resistance within IT ranks, I see that changing as vendors have developed good solutions that are hosted from the cloud. Not everything has to be in the cloud too -- some companies have valuable Intellectual Property (IP) that they don't feel comfortable putting in the cloud yet. In some cases, the amount of data that needs to be fed to the cloud-based systems has created challenges due to latency and network bandwidth constraints.

The definition of "cloud" on Wikipedia is "unconstrained and unlimited". The ERP in the cloud has to be optimized for scale to be able to store and process as much as you want provided you pay. 

Many consumer applications have scalable infrastructures. Facebook never comes back and tells you that you can't post more than 10 pictures in a day, which would make you really upset because you can't post one more photo of your cat in that beautiful pose. Scale is about being able to handle volume without any capacity constraints in an automated fashion. Amazon, Google and Facebook, for example, have a lot of infrastructure automation. Simply put, when their systems sense increasing demand, they are able to automatically adjust computing power and storage space with minimal to no human intervention. 

A scalable cloud ERP system should also sense increasing demands on the processing and storage infrastructure during peak usage times, and the system should automatically adjust and distribute the load accordingly. From a user's perspective, we should be able to go from 5,000 to 5,000,000 transactions per day with the Cloud ERP infrastructure scaling up automatically so that no one notices a difference.

In today's day and age, scalability isn’t a technology issue, but more of an automation challenge. True ERP cloud vendors have infrastructure automation capabilities like those at Amazon and Google. 

You'd be really unhappy if your local power company says, "We are going to shut off your electricity for five to six hours on the third Friday of every month to perform maintenance." You expect ERP in the public cloud to work all the time just like you do your electricity at home. 

When companies run ERP systems in their own data centers, they can bring the system down once a month or quarter to do maintenance work, apply patches and perform upgrades during times when fewer users are accessing the system. However, in the new world of "true" public ERP cloud, if you are depending on the system to help build and ship products 24x7x365, you can't have your vendor tell you that system will not be available during certain times on certain days. 

When a Cloud ERP vendor needs to periodically bring down the system for maintenance, in my mind, that implies that they don't have a true cloud ERP yet. This is because many vendors, even though they're touting their application as cloud based, still operate their applications as if on-prem. They have not yet made their applications cloud-ready and instead offer managed application services on steroids! 

Public cloud ERP should have system availability similar to what's offered by utilities like water and electricity. Some forward-looking ERP providers are already moving in the direction of zero downtime. 

The cost of public ERP cloud deployment should be cheaper because you don't have to buy hardware and spend time installing software in your data center. Since the SaaS implementations typically require configuring the software to your needs and are not customized, the implementation should be faster and the implementation and maintenance costs should be lower. Cloud ERP implementations should reduce overall TCO as compared to on-prem deployments. 

Amazon web services, Google and Microsoft Azure have historically reduced their infrastructure subscription prices because of the economies of scale from the cloud. On the other hand, some ERP cloud vendors tend to increase their subscriptions fees each year or when contracts are renewed. You might wonder why.

Increases in cloud ERP costs are probably due to an inability of ERP cloud vendors to take advantage of the scale of the cloud. As a cloud vendor grows its business, each customer should benefit from the economies of scale by paying less to use that service as more customers are added. However, one can't benefit from scale just by airlifting on-prem applications and slapping them on to dedicated instances. To benefit from scale, one has to have both a scalable infrastructure and cloud-ready ERP functionality.

The next time a Cloud ERP vendor hikes your price when you're renewing the contract, look under the hood and ask why they aren't able to leverage the scale so that costs are at least the same or lower. 

Similarly, cloud ERP should offer price elasticity, more popularly known as "pay as you go". ERP vendors will charge more if your number of users or seats grows, but ask for a price reduction if you have to reduce the number of seats and you most likely won't get one.

ERP data, which generally includes company financials, customer data, Personally Identifiable Information (PII) and IP, should have appropriate security controls in the cloud data-centers. 

Data should be protected:

Also, cloud ERP should have disaster recovery capabilities including appropriate level of backups to protect against system failures and hackers.

Cloud ERP should have relevant Industry certifications, follow established security standards and allow for periodic audits.

How will we get there?

In the area of public cloud ERP, there are mainly two categories of ERP cloud operators. One group uses HyperScale data centers like AWS to host their ERP systems. These ERP systems were originally designed for on-prem operations. The second group includes traditional ERP powerhouses that have developed new Cloud-based ERP systems that are run in their own cloud data centers. This group is mastering the infrastructure automation skills of HyperScalers.

When the two competencies -- HyperScale infrastructure automation and Cloud-ready ERP functionality -- are combined together into one seamless offering, then we will have Nirvana.

1. Software as a Service (SaaS)

2. Scalable Infrastructure

3. Always-ON

4. Lower TCO

5. Risk Reduction / Security