One of the biggest concerns that arise when planning cloud migration is how much downtime your system will experience and to what extent it will affect business operations.
According to the 2017 Cloud Migration Survey, downtime is the biggest challenge of cloud migration, followed by staying within budget, performance disruption, data loss, and security requirements. (Check out the 2017 Cloud Migration Survey Report for more info.)
As part of our Expert Roundup Series on Cloud Migration, we’ve asked six cloud experts to share their cloud migration strategy for minimizing downtime during cutover . How much downtime is too much when migrating to the cloud? Can downtime be avoided altogether?
For previous posts in the series, check out:
Now for their answers to Part 5: How Much Downtime Should I Expect When Migrating to the Cloud?
Kevin L. Jackson Cloud Computing Technical Fellow at Engility Corporation and Founder of Cloud Musings
Kevin L. Jackson is a globally recognized cloud computing expert, independent thought leader for IBM and Dell, and founder of the award-winning Cloud Musings blog. In 2015, he was recognized as a “Top 50 Cloud Computing Blogger for IT Integrators” by CRN.
If you follow the migration strategy suggestions I wrote about earlier in the series (see here and here), there will be zero downtime. Your new cloud process will run parallel to your old legacy process until you turn the old one off.
James Bond Cloud Chief Technologist at Hewlett Packard Enterprise (HPE)
James Bond has 25+ years of experience in roles such as CTO, COO, and Chief Architect. He is a trusted advisor and subject matter expert providing cloud strategy, guidance, and implementation planning to C-level customers through executive briefings and conference presentations. In 2015, James published The Enterprise Cloud: Best Practices for Transforming Legacy IT.
A carefully planned and executed migration should not result in downtime. Industry best practices normally call for individual applications/workloads to be migrated one at a time rather than attempting a massive “move everything” approach. This greatly reduces the risk of unexpected downtime or prolonged migration windows.
New internal cloud services are best deployed in parallel while the existing traditional IT infrastructure remains operational. Once the new internal or external cloud is operational, there are multiple techniques that can be used to copy applications and/or synchronize data from the traditional to the new cloud-based environment.
Many of these techniques allow parallel operation of the new cloud-based application for testing and validation before the actual cutover of end-user activity to the new environment. While the deployment technique for newly developed cloud-native applications is different from shifting legacy applications “as is,” there will still be some scheduled launch windows (e.g. overnight or over a weekend) where DNS, load balancers, and other systems are modified to transparently point end-users to the new application/environment.
Ofer Gadish CEO at CloudEndure
Ofer Gadish, the CEO and co-founder of CloudEndure, is a serial entrepreneur and prolific innovator. He has over 16 years of experience in senior management positions, both in startups and established corporations.
This question exists because there’s a misconception that cloud migration requires significant downtime. With manual processes, this might be true. But automation enables migration with just minutes or seconds of downtime.
There are two parts to a cutover. The first part is the actual cutover, which can be reduced to several minutes by using the right automated migration tool. The second part is the verification of the cutover, this is mostly done manually but sometimes is automated as well. This can take anywhere between minutes to hours, depending on complexity of applications and how good the verification procedures are, and how much automation is involved.
If it’s possible to have just minutes or seconds of downtime by automating the process, I don’t know why any company would choose not to do this.
Matheiu Pierret Cloud Enablement Leader at Cloudreach
Mathieu Pierret is the Cloud Leader for Cloudreach based in Paris. Passionate about IT Transformation and Cloud Computing, Mathieu has more than 15 years of experience in IT project management (he is a certified ScrumMaster and AWS Solution Architect Associate) and manages strategic platforms.
As said previously, our first strategy pillar is to migrate first, ideally via a lift and shift approach. Using the right tooling to automate the move allows you to drastically reduce migration downtime.
From our experience, CloudEndure is the best live migration tool on the market. Once a server is replicated in the cloud, it is kept synchronized in real time, which allows you to move an application with minimal (often zero) downtime. When you need to switch DNS, it’s conservative to say that 10 to 15 minutes of downtime is what you can expect in most cases.
Tom Ray Head of Cloudreach, USA at Cloudreach
Tom Ray has 18 years of experience in enterprise IT and has spent the last 6 years working specifically in cloud computing. After acting as the Global Head of Operations and Cloud Enablement at Cloudreach, Tom now leads the sales and cloud enablement teams for Cloudreach across North America.
Downtime during migration depends upon the application, volume of data, and other factors. I would consider this on a case by case basis.
Jeffrey Kaplan Cloud Computing Strategy Consultant at THINKstrategies, Inc.
Jeff Kaplan has over 30 years of experience in IT/network management, SaaS, cloud computing, managed services, and telecommunications. Kaplan was named among the Top 50 Cloud Bloggers of 2015 by the Channel Company and CRN.
Cloud migration should never cause significant downtime if done properly.
Want to get more cloud migration tips from this expert panel? Check out their answers to this essential question: What Is the Best Strategy for On-Premise to Cloud Migration?