The silver bullet to cloud computing – segmentation
Identifying usage scenarios for cloud computing is a challenge for providers and IT departments alike. Due to that the public discussion on cloud computing mainly gravitates around technology or data privacy. Another blocker really is a row of misconceptions. One of these misconceptions seems to be quite dominant in the range of SaaS – All or Nothing. But even in the other areas, platform and infrastructure, there is no clear view on usage scenarios yet.
Learning #1: Sales Approach
Sell through examples is really what will win in the end. Do not sell by explaining that you got the best technology but come up with usage scenarios (example scenario). Create customer reference stories around great ideas how to use the service. You have to make sure that you create a story explaining the business benefits rather than e.g. the migration story. Even if your migration story puts you ahead of competition it is important to get a picture about the usage scenario into your customers head in the first place. Why worry about migration if I just do not know what to use the service for.
Learning #2: Segmentation is what creates opportunities
For all kinds of public cloud services (from IaaS through PaaS to SaaS) the silver bullet is segmentation. The reasons for that is that unlike in the past where one size fits all was an approach due to limited delivery options/capabilities and lack of scale, the cloud computing approach of “on demand” lowers the entry barrier. To consider cloud computing and build a plan you need to look at your workloads and user groups and analyze them in detail. You will have to come up with a make or buy decision where buy will probably include more than just pure cloud computing. They key piece is to identify workloads that would fit the cloud delivery approach. This could be guided by any of the following areas of consideration:
- Data privacy
- Continuous demand versus peaked demand
- Level of standardization / need for customization
- Cloud as the primary delivery option or an add-on
- Service Levels
- Lead time on demands / lead time on delivery readiness
Let me give you two examples to go a little bit beyond theory. The first example is about a company designing complex products like cars. There are many calculations and simulations to be. You can bet that there is computing power already in place to do the number crunching from variants of the gearbox through to virtual crash tests.
But the reality is that the departments in need of a complex calculation job need to file the request and wait to get a slot granted. Often this could take several weeks just due to the sheer amount of jobs needed. Here cloud computing could help significantly. You might want to look at the jobs and try to understand where you would be able to speed up things by using external capacities in the cloud. Where can you gain flexibility and business agility by combining your internal forces with the computing power in the cloud? It is not about replacing the local number crunching completely but to shave of peaks or react on short-term needs in a more flexible way. As a result you can shorten your lead times on new products while reducing the cost (direct – investment in additional computing equipment; indirect – less idle time or distraction for the teams involved by shorter lead times on results). On top of that segmentation of jobs also allows you to cater for security needs as you can ensure that the top-secret jobs run in-house while less critical ones can run in the cloud.
This example even works in areas usually restricted from using the cloud. In Germany health insurances, amongst others specific business, are not able to generally go to the cloud with PII due to the law (§210 STGB). They might still do it with anonymized data for risk assessment or statistical calculations. Again a usage scenario opens an opportunity in a vertical that seems not be a market at all. Which would be Learning #3: Be creative.
The second example is a straightforward SaaS collaboration example. Most discussions around the pickup of cloud based email or collaboration tools start with a look at the whole user base. By applying this view many discussions are dead in the water from the very beginning.
But by looking at the segmentation of the existing user base and even beyond that, e.g. users without access to IT at all, opportunities to utilize the cloud will appear. It is just a different view that will let you see beneficial scenarios. You could cater for the different needs with the most fitting offering and also cut down cost while not cutting down functionality. The people who need very high reliability and total privacy are being served locally and are happy to pay a premium, the general workforce can be served from the cloud in different forms of usage and the group without access to IT so far may go for a kiosk user model. You might want to think in workloads as well as where you have people working in shifts without an own PC and without the dependency on client software. These users might be totally happy with a web-based cloud standard offering at a lower price point. Also you might want to just use cloud computing for specific maybe temporary usage scenarios. So you create something that might live for some months but instead of buying hardware and software you just use cloud computing for a few users.
If you think about it this where cloud computing starts to deliver real value beyond just cutting costs. It also clearly shows that the so famous “one size fits all” approach of some cloud providers may not the best choice for your enterprise. In the example of Google Apps this was recognized by some partners who have created hybrid solutions and scenarios. So sometime you have to look beyond just the cloud provider to find a solution that fits. But as the mantra for every great cook is “preparation is everything”, you should analyze your need and maybe look at ways of using cloud computing that might not be obvious in the first place.