On May 7, 2013 SAP announced SAP HANA Enterprise Cloud. As the name implies, it is a cloud-based service that allows an organization to move existing (or new) implementations of the SAP Business Suite and SAP NetWeaver Business Warehouse, powered by HANA, off their own servers and into SAP’s massive data centers. Why would an enterprise want to do this? The short answer: Speed, power and the benefits of cloud computing without the disruption of replacing existing on-premise solutions. Speed and power come from HANA, adding visibility and agility to the business by enabling decisions to be made in real-time with volumes of data that were inconceivable just a short time ago. Cloud computing lowers cost and adds elasticity, allowing capacity to stretch as your business and your need for data grows.
This announcement builds on the progress SAP has made in bringing HANA to market, first with analytics and then applications powered by HANA, including both the Business Suite and Business One. Because HANA is essentially enabling technology and it is technologists that deliver the message, the significance and potential for business value is difficult to convey. This newest announcement is focused on the cloud, but without first understanding the power and potential of HANA itself, it loses its punch. So let’s backtrack a bit and highlight what HANA brings to the party.
SAP HANA is the brainchild of SAP founder Hasso Plattner and is often referred to as Dr. Vishal Sikka’s “little girl.” Professor Plattner is Chairperson of the Supervisory Board of SAP AG and Dr. Sikka is Member of the Executive Board of SAP AG, Technology and Innovation. Both men are brilliant technologists. They speak quite eloquently about the possibilities created by this breakthrough technology, but often that eloquence is lost on business executives. The non-technical businessperson cares little about row versus column processing, in-memory databases, massively parallel processor arrays (MPPAs) or multi-threading.
However, they do care about speeding up processes like Material Requirements Planning (MRP) in a manufacturing organization or forecasting demand to optimize restocking the shelves in a retail environment. These and other “batch” processes are the Achilles heel of most ERP systems. Even as hardware innovation continues to accelerate, because they are “batch” and not real-time, they are run overnight or over the weekend. Between “runs” things change and therefore decisions get made with something less than a full and accurate picture of the world.
Benchmarks from early projects have proven that HANA can reduce these run times from hours to minutes, or even seconds, without constraining the amount of data processed. In fact doors are opening to bringing in volumes of data that were previously inconceivable. However, having been constrained for so many years, it is often difficult to look beyond current real and perceived constraints and understand the possibilities. While a business executive might not really care how this is accomplished, understanding some of the concepts to a certain extent might help business executives and their IT staffs see the possibilities. Without that vision, businesses will never make good use of the technology.
Parallelism versus multi-threading is one of those concepts. Most computers today appear to be able to do many things all at once. After all, you have multiple users logging into ERP and CRM and other applications, and you can record an inventory transaction at the same time you are entering a new customer quote or matching invoices to cash receipts, right? Not really.
A processor in a computer can really only do one thing at a time. But it breaks each of those transactions down into minute elements and strings them together into a “thread.” If the processor has 3 tasks to do, it does one element from one thread, then an element from the second thread, then one from the third before it cycles back and picks up the second element of the first thread. Because these elements are so small and this is happening so quickly, it appears as if it is doing all three things simultaneously, when in fact it is doing only one thing at a time. Think of it like going through a turn-style. Only one person fits through at a time.
Going massively parallel allows you to throw a large number of processors (or separate computers) at a task (or multiple tasks) to perform a set of coordinated computations in parallel, like adding hundreds or thousands of turn-styles. In order to do this, you might have to modify or even redesign the subway station or the football field where the turnstyles function, so it is not as simple as it might appear on the surface. But the net effect is to increase the throughput by several orders of magnitude. Yes work needs to be done to the application before it can take advantage of the power of HANA, but the potential for adding throughput, and therefore speed, is nothing short of amazing.
Of course this is an over-simplification, but hopefully it helps the non-technical business person see the vision of the kinds of massive volumes of data and computations that can be handled quickly and efficiently with HANA. One goal of SAP HANA is to eliminate the need for “batch” processing. It reached this goal by increasing the speed at which computations can be completed. According to Prof. Plattner, “There is no batch any more. After 40 years of trying to kill it, batch is now a monster of the past. We can do this now.”
While speed is great, speed for the sake of speed alone adds little business value. What is more important is the ability to make more real time decisions.
While computers and enterprise applications have opened new doors to new possibilities over the past several decades, most of us have also become accustomed to being constrained largely because our ability to generate and collect data has far exceeded our ability to process it effectively for decision-making. We need to break through those constraints in order to create the vision of possibilities:
- We run MRP weekly. We plan and schedule assuming infinite capacity. What if we could run a full ERP the moment a new big order came in to see how it would impact the finite capacity throughout the enterprise?
- We forecast demand by product line, not individual products, or by region, not by customer. What if, instead we could forecast by product for each customer by individual region?
- We put a plan together to replenish shelves in a retail store mid-day based on sales last week or last month or even last year. What if we could re-plan at noon based on sales that same morning?
- We analyze the performance of trade promotions once the program is completed. What if we could see the impact on profitability in real-time throughout the life cycle of the promotion?
Overcoming these perceived constraints requires a new way of thinking. At SAP, that new way is called “design thinking.” This concept was not invented by SAP, and SAP is by no means the only company that is doing it. But SAP has certainly embraced it fully. Virtually every employee directly involved with customers and/or products has been trained.
Design thinking is an important concept in unlocking the power of HANA. The SAP Services organization is engaged to conduct a comprehensive assessment, including a design thinking brainstorm session. The goal is to determine which solutions will have the highest performance impact as a result of moving to SAP HANA Enterprise Cloud.
Moving to the Cloud
This brings us to the real meat of the announcement: managed services delivered via the cloud. Perhaps the best way to describe SAP HANA Enterprise Cloud services is to point out what they are not:
- This is not an option for companies to run applications powered by HANA in their own private clouds. It is really a private cloud for the customer managed by SAP. This was a purposeful decision on SAP’s part since the objective is to make the solution truly “elastic.” While this term may be common in technology circles, it is less so in the business community. Essentially, it means the customer is never constrained by hardware limitations. Data center configurations will expand (transparently to the customer) as more computing power is required.
- This is not Software as a Service (SaaS). Customers will bring their own licenses to the party. There will be no multi-tenancy (multiple companies sharing a single instance of the software). Right now the offering includes all of the SAP Business Suite except for SRM (which has not yet moved to HANA), and/or SAP NetWeaver Business Warehouse, any or all of which must be powered by HANA. But SAP also welcomes other applications that might be developed using the HANA platform or existing applications that will first need to undergo a transformation whereby they will be powered by HANA.
- This is not a public cloud like Amazon Elastic Cloud Compute (EC2) where any and all different types of applications can run. This is an environment specifically for HANA-powered applications.
These are managed services, where customers outsource day-to-day responsibilities for managing some segment of their solution to SAP. As part of the ramp-up process for any prospect or customer, the SAP Services organization will likely perform a comprehensive assessment, delivering advisory services to determine which solutions will have the highest impact on performance – “likely” because this is not a mandatory step. SAP pitches a design thinking session and this generally results in at least one project and may result in a whole bevy of projects or a sustained rollout.
Once a project, or projects are defined, SAP Services provides migration and onboarding solutions. For an existing SAP customer this might include bringing the application up to the latest release. Remember an application can’t just be dropped onto the HANA platform. SAP has spent a lot of time and effort “powering” the Business Suite with HANA. The fastest way to take advantage of this will be to come up to the latest release.
The ease or difficulty of this upgrade (and migration) will largely depend on how far behind the customer is and how customized its solution is, and therefore the range of difficulty will be quite broad. The prime candidates for SAP HANA Enterprise Cloud, at least initially, will be large enterprise customers, and they are most likely to have customized the application extensively. Small to midsize enterprises (SMEs) and very new large enterprise customers are less likely to require extensive customization because configuration options and the ability to tailor without programming have evolved dramatically over the past several years. But more mature implementations will probably need to overcome this hurdle.
If this, or other hurdles can be overcome, then the benefits of moving to the cloud can be very significant. The benefits are largely based on the elasticity noted above and cost related factors. The customer should never have to wait for new hardware to be evaluated, selected, delivered and configured – which is exactly what would happened if they continued to operate from their own data centers.
Mint Jutras research finds that cost and upgrade factors are perceived as the most important benefits of a SaaS offering (Figure 1). Given the tendency for many to equate cloud and SaaS (even though they are different) there is no reason why we can’t extrapolate these perceptions to this offering. Survey respondents were presented with five general categories of benefits of SaaS and asked to sequence them in order of priority (with 5 being the highest). The numbers shown in Figure 1 (the mean average response) denote the relative order of importance.
Figure 1: Relative Importance of Benefits of SaaS
- Cost factors: 3.59
- Upgrade issues: 3.09
- Support of distributed environments: 2.90
- No hardware purchase/maintenance required: 2.75
- Less IT expertise and staff required: 2.67
Source: Mint Jutras Understanding SaaS
While the upgrade factors can be directly applied in terms of the hardware, they will be a little different in terms of SaaS versus cloud services that are not based on a multi-tenant model. Cost factors are also a bit different. Customers will not experience savings in terms of software licenses with the “bring your own license” model. However, there should be some savings in trading hardware costs for services.
But the real cost factors come into play when you measure value. Recognize that you are making a quantum leap forward in terms of taking applications to an entirely different level in supporting real-time decision-making. Remember those “What if’s?” listed earlier? Without this type of option, most companies simply would not be able to afford this move, and even if they could, they might find it difficult, if not impossible, to keep up with the level of hardware and technology innovation being delivered today.
SAP HANA presents a whole new world of opportunity for speed and power, limited only by the customers’ ability to see the vision of what is now possible. It is most important to understand the potential and identify your own possibilities for innovation. These opportunities for innovation may be staring you in the face. Or you may have to apply some design thinking: dig deep into existing processes and identify those problems you have been living with for so long you might think they are unsolvable.
In the words of one early adopter: HANA solves problems that were deemed unsolvable in the past. As you identify these opportunities, the next question will be, “Can I afford the solution?” With the introduction of SAP HANA Enterprise Cloud services, the answer is far more likely to be a resounding, “Yes!” SAP HANA Enterprise Cloud also adds the ease of consumption of the cloud at an affordable incremental cost.
So once again, why would an enterprise want to do this? The short answer is also the best answer: Speed, power and the benefits of cloud computing.