
Know What to Look for in Cloud Hosting Data Centers
Submitted by Michael Ray, Solutions Architect for VAZATA.
With all the vast new options that Cloud services introduced in 2011, flexibility and scalability were two areas abuzz with conversation. The fact is operations teams must remain as diligent in designing Cloud architectures as they would a traditional approach. It can be dangerously easy think of the basic reliability issues as ‘somebody’s problem other than yours’, but the fact is, you need to ask the same questions about reliability and scalability that you would ask a traditional provider.
Planning
Determine the expectations for your infrastructure first. Get input from your colleagues and make it a well thought-out collaborative effort.
- Which Web applications are mission critical and require 100% uptime?
- Are there back-end applications that can be down for a few days in the event of a disaster?
- What are the costs of downtime or data loss?
Provider
With your objectives in mind, you can now start looking at providers. Start with a wide selection and narrow it down – five is good to start. For all but the simplest projects, make sure to start with a conversation with someone in person. Discuss initial pricing at this stage so you have a better idea of the market.
Locations
The first step in determining location must be based on answers to the following questions:
- Does your application require certain latency or performance guarantees that will be impacted by network placement?
- Does the facility meet Tier 3 or Tier 4 standards as set by the Uptime Institute?
If all of your users or visitors are in Seattle, it probably doesn’t make sense to put your datacenter in Miami. Applications requiring 100% uptime must be hosted at more than one location, and subsequent locations must be geographically diverse. If your application requires 100% uptime, you should take your analysis of locations one step further.
- Are there any predictable events that could impact multiple locations?
The best news is with Cloud the backup capacity is economical. You may have only a few servers—or none at all—running in the backup datacenter, with the ability to spin up more in case of an event. Ask providers what they recommend for disaster recovery. Then make an informed decision.
Network
Network connectivity is vital. Questions to pose may include:
- Is the provider connected to multiple “Tier 1” Internet providers?
- What steps does the provider take to ensure that there aren’t single points of failure in their network access?
Data and Monitoring
By now, you should have a good idea of the questions you need to ask to ensure your Cloud provider is the one for you. But one more question goes back to the planning stage. Do you have a documented, regularly tested process for failovers?
- Where is your data stored?
- Is your critical data still accessible if your primary datacenter goes down?
- How long will it take to transition to the DR Datacenter – and can that time be improved?
On the subject of monitoring – ask these questions:
- How will you know if a critical service is offline? If you don’t find out about an outage until you arrive at work on Monday morning, all of your disaster recovery plans will be compromised.
- Are all critical systems monitored?
- Do the people getting the monitoring alert have a documented way to engage the disaster recovery process and communicate the status?
- Do all critical personnel have contact information and details of processes?
The Cloud is perfect for companies looking to deploy scalability and reliability for websites and web applications. Just be sure to plan well and avoid becoming a media statistic, losing customers and customer confidence as a result. Think ahead, make it a team effort and most of all, ask the right questions right from the start.