Ok, I'm actually finding some reasonable use cases for 3rd party cloud solutions. I still don't think it makes sense in instances where multiple large applications and lots of integration's are involved but for "point" solutions with few integration needs it may make good sense. The cases I am thinking about (outside of family :) ) are development tools/process related. With staffing so slim, it is pretty risky having only one person that knows how all the development and operations infrastructure is built, installed, setup and maintained. Having someone else deal with most of that means that staff have more time to focus on actual development. As is, maintenance of development tools/processes is falling to the wayside more and more for lack of time - that increases risk that something could go very wrong resulting in loss of assets or substantial disruption to systems/processes. I will say that a cloud solution isn't a silver bullet though - it has its own risks like provider outages, network disruptions etc which can have a large negative impact on development processes too. Which risks are more likely and have the higher negative impact? Hard to tell right now but worth thinking about.
Anyways, use the appropriate tool for the job..
-------------------------------------------------------
This post is a bit of a general complaint against some people in IT and management. It seems that there is an increasing tendency for people to request "cloud" applications or make requests to convert existing applications to "cloud" applications. Why should this concern anyone? I find it concerning because these requests are made for a specific solution instead of having a specific need or requirement where "the cloud" is the best solution. The decisions behind any IT solutions should be made to solve specific problems or meet specific requirements. Technology decisions based mainly on some research companies "quadrant" leads me to
- wonder whether the solution is really a good fit for the organization
- concern regarding the longer term viability of a solution
- concern whether a project is setup for failure from the start
- concern that the total/on-going cost will be wasteful of money and/or other resources
The type of situation that leads me to finding a lack of in-depth understanding includes someone trying to justify a cloud solution by saying it will save money but can't back that with any further details. If the only further justification is "we don't need to by hardware" then you are likely in the same situation. If the application/service is a mission critical application then you may end up with a really big problem.
If someone wants to talk about cloud solutions with me then I expect them to be willing to start the discussion with the following items at a minimum:
- Security requirements
- application access
- data sensitivity
- User base
- size
- locations
- types of users
- public
- internal only
- etc
- Data sources, locations and integration types
- Performance/capacity requirements
- Availability requirements
- Disaster recovery requirements
- Compliance/auditing requirements
- Potential internal staffing needs
- training
- developers
- operations/support personnel
- Potential network requirements
- cloud types
- 3rd party - external cloud / hosted solution
- internal cloud
- hybrid
- Potential maintenance logistics/coordination
- If you require specific availability windows unsupported by the 3rd party then what?
- If you have maintenance windows for integrations/data sources then how is that handled/coordinated?
- Help desk / communications logistics
- If there is a problem, how is it reported?
- How are users notified?
- Monitoring
- how is it performed? Can you leverage existing infrastructure?
- Who performs it?
- what are the boundries? If there is a performance issue, how do you identify whether it is a cloud issue or a data source integration issue?
- Contracts
- Do you have adequate legal counsel to help in evaluating contracts so you are not left high and dry if a problem of any sort comes up?
- hosting company goes out of business
- hosting company raises rates beyond what is affordable
- hosting company can no longer meet your needs
- hosting company not adaptable to changing technology needs
- Cost model
- Are you able to account for all changes required by your organization and still determine that there is a real cost savings?
- Can you get all your data out?
- What format is your data available in?
- How transportable is your application/service to some other cloud vendor? Do you have to rewrite your application from scratch?
No comments:
Post a Comment