One of the major challenges when looking into cloud architectures is the organizational/process shifts required to make the transition. Technology refreshes seem daunting but changing organizational structure and internal auditing/change management can make rip-and-replace seem like childâ€™s play. In this post Iâ€™ll discuss some ways in which to simplify the migration into cloud (thereâ€™s a vaporize your silos pun hidden around here somewhere.) For more background see my post on barriers to adoption for cloud computing (http://www.definethecloud.net/?p=73.)
There are several ways in which companies can begin the process of migrating into the cloud and preparing their IT organization for the transition. These same principles can be applied to each cloud architecture type and to the concept of data center virtualization. In order to properly the first step is defining a cloud strategy:
Cloud strategies will be unique per company but should include the following (not an all inclusive list):
- Migration goal. This should be based on the ROI calculations driving the move to cloud.
- Defined timeline complete with checkpoints and dates. Timelines for this type of transition will most likely be multi-year, ranging from 3-10.
- Type(s) of cloud architectures being considered and used.
- Staged approach outline defining the migration flow including details of which apps will be moved when where.
- Risk assessment for each piece of the plan and the plan as a whole. If the migration could end up costing you more than the ROI will buy you itâ€™s time to rethink.
With strategy in hand you can now begin planning the organizational migrations that will be required to make the move to cloud computing. The first step is a history lesson (you should never move forward without first looking backward.)
Our data center has a sordid history which is often forgotten because it all occurs in a short lifespan of about 30 years. We scaled out from main frames to commodity hardware, then up to dense hardware, then virtualized to repair utilization problems and ended up in our current mess.
During that process weâ€™ve built large technology silos and tailored our organizations to them. Weâ€™ve actually built our processes around the technology mistakes or oversights weâ€™ve made. Weâ€™ve architected siloed organizational structures to marry our departmental or technological silos. Breaking those silos is not an easy task but it is key to moving forward.
You can not and will not break out of technological silos without first breaking organizational silos.
Organizational change cannot be an instantaneous thing in most companies. Even when it could be, itâ€™s best to err on the side of caution. Assess your goals, reference your strategy and plan the changes to break the silos and bond the data center teams.
Some short-term suggestions for long-term results:
- Place Network and Storage teams under common management.
- Do this as close to the administrators as possible. Regardless of technology the storage network and traditional data network will converge. By consolidating management you will begin to build cohesion and knowledge base prior to technology changes.
- Cross-train your MVPs.
- Training is key to developing and maintaining the best, but cross-training is key to moving forward. Take the best from each silo and train them across a second silo, possibly even going so far as to having them temporarily swap roles. Each contributor needs a strong understanding of what their cross-department colleagues need.
- Promote ingenuity.
- Most organizations will have staff adverse to any change because they donâ€™t understand it and or want to learn something new. Many people feel they can cling to the past as long as possible and will do quite a lot to do so. Encourage innovation and education within your organization, promote the embrace of new ideas and technologies. Test new concepts in limited environments to ensure they work for your company.
- Analyze budget process.
- If the silos in your organization each own their own budget consider moving budget further up the chain. I/O consolidation is a great example of this. If you choose to consolidate networks which budget is responsible? Prepare your budgets for data center structural changes.
These are just a few broad concepts to expedite the planning of your organizations move to the cloud. Overall cloud will be no more of a rip-and-replace of your data center than a rip-and-replace of your organization, itâ€™s as gradual a progression as you plan, so plan it well.