I love politics, it provides a daily real life Dilbert cartoon. Whether it is Carter, Reagan, Gingrich, Pelosi, Bush (either one), Clinton, or Obama, too many “winners” are equating victory with a mandate, when in reality it represents total disgust with the incumbent. Even more disappointing is that in the most recent mid-term elections, many of the moderate incumbents were replaced with their more radical opponents. The result is a congress that is moving even more to the edges while the majority of Americans are closer to the center. I guess it is good news that there are now enough radicals on the right to balance the radicals on the left, but dang it hurts when the winning side tries to stuff its radical ideas down our throats.
What does this have to do with the Cloud?
You’ve been trying to convince your boss to move to the Cloud, and she has directed that you do so. Finally, you think, sanity has arrived and you will be able to make the sweeping changes you know your company needs to save money and stay competitive. You think “Victory is mine!”
Not really. Your boss is fed up with the status quo, and wants change. You have sold her that you can provide that change with positive benefits to the organization. Just like Congress, you have one opportunity to show that benefit or face her wrath at your next performance review, or your exit interview.
In a recent blog entry, Walk, Don’t Run, to the Cloud, I talked about some of the characteristics to look for in that first project. In that blog, I suggested that you look for an application that is not critical to your operation, doesn’t have strict security requirements, has reasonable availability and performance requirements, can be moved to the Cloud without changing a lot of internal processes, and which can show some real financial benefit.
Some potential candidate applications, although knowing your environment will probably suggest others to you:
- Test and development.
Test and development is often a good first candidate. It has all of the right characteristics: the data is usually test data so its security isn’t a real concern, if it is down for a day you don’t risk your business, and other than for the final stress testing, performance isn’t a serious consideration. The people responsible for the test and development are usually the more computer savvy people in the organization so they can quickly tell when there is a problem and usually how to deal with it, they like to learn new things, and you want them to learn about the Cloud anyway. Where does it go in the Cloud: usually PaaS (Platform as a Service) in the Public Cloud.
- Commercial applications you are running like e-mail, CRM, financial applications.
Many of these applications already exist in the Cloud, either from the software vendor or some independent CSP (Cloud Service Provider). I suggest you don’t make any application configuration changes at the same time, as that will change some of your internal processes. Get the application in the Cloud first, then make the related changes. Where does it go in the Cloud: Software as a Service, usually in the Public Cloud.
- Maybe some of your own applications.
If you have an application that you developed and still maintain that is largely unconnected to other application solutions and almost entirely accessed over your internal network, it might make a good candidate for your first Cloud application. Again, your people understand it and there should be minimal process changes to the users of the application since they will continue to access it over their network is the same way. Where does it go in the Cloud: probably IaaS (Infrastructure as a Service) in the Public Cloud or Private Cloud depending on data security and performance requirements.
Some not-so-good ideas to pursue first:
- Disaster Recovery.
It seems so obvious – use the Cloud to provide your disaster recovery capabilities. It probably can, and at substantial savings to what it would cost to do it yourself. It also provides the potential for addition DR capabilities. While you could do it one application at a time, you will only see real financial and operational benefit when all of your critical applications move. That requires that you solve all of the security, performance and availability requirements for all of your critical applications at one time. If you add additional DR features, you have now complicated the financial measurement: what are those new capabilities worth compared to the prior environment? Am I really saving money? You never really know until you have that disaster, and that is hard to orchestrate in advance. There are two many facets to show a quick neat success.
- A brand new application.
Again, it is easier to create something to run in the Cloud originally than to move it to the Cloud later. However, now the success of your Cloud initiative is tied to the success of the new application. Problems in the application may be perceived as “the Cloud doesn’t work” and problems with the Cloud implementation may impact the application deployment schedule. When you are done, you really don’t have any good way of saying “see it worked and I saved you this much money a year.”
As I discussed in the earlier blog, the key is planning.
- Find all of the stakeholders. Stakeholders are not just in IT, but include any users of the application or its output, including customers and the public.
- Talk to the stakeholders and make sure they understand what is happening, the schedule, and any impact to them.
- Work with the selected CSP to plan the migration to the Cloud.
- Develop an acceptance test, and test the transition before going live.
- Have a plan to communicate any postponement of going live until you meet the acceptance test. And do postpone until you are satisfied with your test results. Better to be a little late than fail.
- Have a plan to gracefully back out of the Cloud in case something serious is discovered later.
- Keep talking to your stakeholders.
If the first Cloud exercise is a success, you will have a far easier time as you take your business into the Cloud and reap the benefits.
Unlike politicians, don’t disappoint your stakeholders.
The last word:
Will the politicians learn their lesson this time? It is hard to imagine they will since they have been hit by this proverbial 2×4 on the last three mid-term elections and about half of the last dozen presidential elections and missed the message. The presumptive Speaker of the House, John Boehner, actually said that he got it: they didn’t win, the other team had earned the voters’ wrath. I applaud him for saying it, but Representative Boehner is a politician and his lips were moving at the time. President Obama doesn’t get it. He said that the “people didn’t get it” and is blaming we the people for being too dense to understand his policy. So I have little hope from him. Personally, I expect more of the same, and the need to give them the same lesson again in two years.
Keep your sense of humor.