When I look at the U.S. market, cloud computing has probably been the biggest topic since 2010. Small, medium-sized and large organizations alike are trying to take advantage of the various “cloud” offerings from different vendors—either because they expect cost reduction, more flexibility, or both. Let’s take a closer look at cloud computing to see what the term means, how it fits into the SAP world and how we at itelligence have prepared for it.
Currently there is a great deal of confusion around what cloud computing actually means, in large part caused by vendors defining their cloud offerings differently, most often to best fit traditional service offerings or setups.
The U.S. National Institute of Standards and Technology (NIST) has defined cloud computing as “a model for enabling convenient, on-demand network access to a shared pool of configurable computing resources (e.g., networks, servers, storage, applications, and services) that can be rapidly provisioned and released with minimal management effort or service provider interaction.” This definition is part of a document that is well worth reading, available here.
The key criteria for a service being considered a true “cloud service” are:
• On-demand self-service
• Broad network access
• Resource pooling
• Rapid elasticity
• Measured Service
Now that we know what a “cloud service” looks like—or at least what it should look like—let’s consider the different definitions of the various cloud types.
First we have the so-called “private cloud,” which consists of an environment that is supporting a single organization. Such an environment has to meet all the baseline criteria, but most service providers would refer to a simple environment utilizing server virtualization like VMware, Zen, AIX or HP/UX as their “cloud.”
Then we have the “community cloud” which is pretty much the same as the “private cloud,” with the difference being that multiple organizations with the same requirements share such an environment.
A “public cloud” service is provided to the full public, meaning individuals or organizations that don’t share any common requirements.
Last, there is the “hybrid cloud” which consists of any two or more of the just described “cloud” types.
Confusing? I would agree, and ultimately, every service offering that follows the basic criteria can be called a “cloud service.” But it is not only the different “cloud” types causing that confusion, but also the way organizations loosely (re)define cloud computing criteria so they can call their own offering a “cloud service.”
Now that we know what a true cloud service should look like, let’s see how cloud-computing requirements fit into our SAP world.
This is the capability to change (add and remove) computing power by the user/customer as needed. For example, an additional application server can be added to the ERP instance to have more computing power for the month end closing, or it could be the automated provision of more computing power in a virtualized environment by flexible CPU assignment. It also can be the automated, on-demand provisioning of any SAP product for testing purposes or an instant snapshot copy of a production system. Key with these criteria is that it is an on-demand, mostly automated process.
Broad network access
This involves the need for sufficient bandwidth, and that the “cloud” environment actually supports different end-devices (PCs, tablets, phones etc.). In an SAP environment, broad network access is basically a given since the application itself supports this functionality.
This criterion is fulfilled by every organization using any kind of server and SAN virtualization. Development, quality assurance and production systems share a pool of disks (typically a SAN) and CPU resources of a virtualized server or server clusters. Currently, the sharing of memory between SAP instances or products is not yet possible.
This ties into the on-demand self-service. It requires resources to adjust automatically or as requested in a quick way—and the elasticity is supposed to be “endless.” In a virtualized environment, this elasticity is always given for the CPU power, but when looking at a typical SAP landscape, this would not be sufficient. For example additional application servers that automatically start in a short time frame would be required.
Again, this is something that every virtual-server software does on its own by monitoring CPU, memory and disk IO utilization of the clients’ virtual machines.
As you can see, the scope for fulfilling the “cloud service” criteria can be very broad or can be very narrow, and we should not expect all SAP cloud service providers to be able to provide everything that one can imagine as a readily available cloud service.
For most enterprises, the automation component is not ready for prime time. In many cases, SAP system landscapes are simply too complex to be ready for fully automatic deployment, short-term provisioning, or even resource elasticity when it goes above and beyond pure CPU horsepower. (Just think about the complexity of the communication between SAP products).
At itelligence, we started the discussion around globalization, virtualization and fast provisioning many years ago, long before the “cloud” began enjoying its recent popularity.
As a long-term partner of SAP, we sometimes get insights into SAP strategic direction from a technology standpoint. Virtualization has been one of the directions that we saw around the time VMware established itself in the market. It has been and still is a great concept to better utilize server capacity and ultimately allow us to provide our services at a reduced fee to our customers, thanks to a better use of landscape capacity. Nowadays, virtualization (with VMware and other products) is a fully integrated part of our standard offering. Besides the server virtualization, we also implemented storage virtualization products that allow us to utilize different SAN types and vendors in an environment that automatically provides capacity to our customer’s systems and allows easy / automated load distribution in the storage department.
While currently being a hype, there is a market for more flexible provisioning, better resource utilization and lower TCO and many organizations already have implemented parts of a cloud strategy, are on the way to do so, or simply do not see the benefit for their organization yet. For a large service provider like itelligence, the optimization of operational processes, better and faster services, and lower fees for our customers have been our key objectives years before “cloud” became a trend. For us, “cloud computing” isn’t a marketing ploy, but a way to describe how we have approached our business for many years.
In the following weeks we will have a closer look at:
• On Demand – Quick and automated turnaround of requests for landscape adjustments like application server provisioning and automated test system provisioning
• Current virtualization technologies: CPU, Memory, SAN, network
• Measured services offerings: SAPs-based agreements and reporting
• Amazon/SAP Cloud offering in the USA: An interesting approach