At Gartner IT Expo, Ephraim Baron spoke about emerging trends:
Cloud computing is not a single technology, it’s dozens of them. Understanding how to proceed isn’t simple. In this session, we examine 5 emerging cloud technologies and discuss how to get significant value from them while avoiding false steps. If you’re interested in cloud migration, container management, cloud service expense management, cloud managed services, or site reliability engineering, this session should be on your schedule.
Set the Stage
- Cloud is just delivering what IT outsourcing (ITO) promised
- ITO just transferred the problems and created new challenges
- On-Premise is heavily customized and fragile
Cloud is Transforming business. It’s something we’ve seen before. The AC vs DC electrical current war with Edison vs Tesla. Someone will win and it won’t be on-premise data centers.
Quote: by 2025 80% of corporations will shut down their own data centers
Quote: Cloud is a mess, not a strategy. It’s a means to an end.
Cloud needs to be a business discussion. You’ve decomposed IT into consumable services.
- Cloud assumes failure so you have to architect for it
- Cloud will NOT save you money…. at least not right away
This entire platform has a number of technologies. Think of it as a set of services you can use. Storage, CPU, authentication, caching, etc.
This is the process of moving applications and workloads from on premise to external cloud services.
- Set busines objectives
- Determine workloads
- Plan how to best migration
- Execute the migration
- Learn from what you’ve done
How to achieve success:
- Create a cloud plan
- Redesign or lift and shit
- Knowledge of your destination(s) – clouds are not all the same
- End to end cloud expertise
- Plan, build, run, automate
- Automate is where you finally get the
Containers allows apps to be easily deployed and managed both on-premise and in the cloud. Container management systems enable continuous integration and delivery. This gives you a whole lot of options on where and how to run your workloads
There’s a lot of interest in this but expertise is scarce. That’s why it’s still pretty far back on the hype cycle. Any book when published is already out of date
Best Practice: You need someone just trained on the latest and someone constantly learning
Cloud Service Expense Management
This is VERY important. It’s the practice of reviewing and reconciling the charges for services provided by external cloud service providers.
- Many organizations don’t understand their external cloud consumption and the associated expenses
- IT leaders need a strong understanding of their external cloud expenses
- Over-provisioning can be costly. How many engineers
- You need to align expenses to needs
- Monitor and manage cloud expenses. Bill the consumption to the business unit. Let them see it
- Rapid access to resources increases their consumption
- If a developer has 6-10 dev environments then you can run up the bill pretty quickly
- Cloud changes consumption patterns
- You can ask a lot of questions and use as much resources as you want
- Developers do not decommission capacity
- Cloud increases cost exposure
- You don’t secure it correctly then crypto-currency miners may take advantage
- Yes, it’s happened
Bottom line: you have to actively manage your costs
Cloud Managed Services
IT Service offerings that provide for day to day management of cloud service environments. This includes configuration, performance management, etc.
Cloud managed services are different. You can’t just port to cloud and it all works the same.
Best Practice: Automation is key. It increases reliability and reduces costs
Site Reliability Engineering (SRE)
This is at the very beginning of the hype cycle.
It’s a collection of processes and principles to operate systems at scale. Google create it and wrote the book on it a decade ago. Think of it as the how where DevOps is the what. So it’s a more mature form of DevOps.
SRE seeks to balance stable environment with the ability to make change.
Moving to SRE means change. The bullets below are the from the old to the new.
- Not manage infrastrucrure but manage app service
- Break fix to permanent fixes to issues
- Measure server uptime to Measure application uptime
- Monitor infrastructure to Monitor KPI’s