Connectivity is just the beginning. Amazon Web Services and Deloitte are committed to empowering, optimizing, and integrating digital solutions that give businesses in every industry the power to adapt and thrive.
How a Large Media Company Designed and Executed an Exit Program from 8 Global Data Centers to the Cloud
BrandPosts are written and edited by members of our sponsor community. BrandPosts create an opportunity for an individual sponsor to provide insight and commentary from their point-of-view directly to our audience. The editorial team does not participate in the writing or editing of BrandPosts.
By Somen Chanda & Manzoor Ismail
In 2018, a leading US-based global provider of business information services, with products including highly specialized information-enabled software and tools for legal, tax, accounting and compliance professionals, combined with the world’s most global news services organization and spun off a line of business that required it to exit from eight data centers around the world.
The data center exit required the migration of approximately 300 services and customer-facing applications (10,000+ servers) over 24 months. Premiere AWS consulting partner Deloitte, in partnership with AWS, led the migration.
The scope entailed migration of 360+ application environments, 5,000+ servers, 37,000+ databases and approximately 80TBs of data as shown below.
As an AWS migration specialist, Deloitte was engaged by the organization to partner with AWS Professional services to:
Analyze applications portfolio for suitability and placement. Based on this analysis:
35% of the applications were assessed and migrated using a lift & shift approach
40% were re-platformed
20% were refactored
5% were relocated to another CoLo facility due to unsupported technology and application dependencies.
Define a detailed target state design for each application
Devise a multi-phase plan for migration, which spanned 4-6 environments and multiple data centers across four continents. Primary consideration of the wave plan was based on technology and application inter-dependency, blackout & freeze periods, team readiness and product release schedule.
Coordinate migration efforts with business, application, IT shared services teams, and AWS/AMS
Formation of the migration factory and execution of all migrations
Seamless hand-off of operations to the organization’s support teams
Through our partnership Deloitte, AWS Professional Services, and Amazon Managed Services (AMS) partnered, shared responsibility and ownership on various aspects of the migration program.
Assessment and Migration Approach
Cloud migrations led by DATA CENTER exit or Enterprise application modernization strategy are generally complex and produces many challenges. The migration approach, plan & sequence of activities are key to the success of these programs. Deloitte’s approach to this migration and for other similar programs is to start with Discovery & Analysis to assess the cloud suitability, create a migration path and roadmap. Followed by migration execution which included activities to perform application deep dives, target state architecture, pattern development, provisioning & deployment and cutover.
As shown in figure 2 above, phase 1 included discovery of key application and infrastructure attributes to determine application cloud suitability, migration path and to create a migration wave plan. Phase 2 was focused on migration execution which included pre-migration, migration, and post-migration activities. Pre-migration activities were conducted to collect detail application details, create target design and migration plan. Migration activities involved scripting, provisioning of infrastructure, deployment, data migration, testing and cutover planning. Post-migration activities involved creation of runbooks and on-premises server decommissioning.
Key Challenges to Migration
During the data center exit there are many moving parts such as dependencies, blackout or freeze periods, cloud suitability, license negotiation, etc. There are various stakeholders, multiple vendors and third parties that required to collaborate and coordinate. There are several factors and constraints need to be handled; some of the challenges that needed to be treated uniquely are as follows:
Aggressive and non-negotiable data center exit timelines: As many data center exit programs deadlines are immovable, it was extremely critical the exit happens on time to avoid financial impacts. Application owners had competing priorities; while the data center exit was ongoing, product releases had to be managed, QA and testing needed to focus on releases, while data center exit testing was required. To mitigate these situations, it required extensive planning with granular tracking to make sure that the migrations are on track with their activities to avoid resource constraints.
Need for hybrid setup: Over a period, organizations will have components and workloads spread across multiple data centers. Certain decisions in the lifetime of an organization such as M&A or cost optimization would result in a need to exit one or more of the data centers. This in turn results in in the hybrid environments that we see across many companies today. In most cases, organizations that formulate a holistic strategy to IT growth that accounts for M&A activities lower their risk from these hybrid scenarios.
Migration of large data sets between On-prem and Cloud: Client had 80 TB of data across 37,000 databases running Oracle, MSSQL, MongoDB, & MySQL. Ensuring data is migrated from data center to cloud with no loss of data and minimal impact to client is quite challenging. During migrations, due to time constraints during cutover, using the right technology for data transfers and synchronization between data center to cloud is key. This requires engaging key stakeholders for real time decision making and taking calculated risks (for example what tool or method should be used for synchronization, how easily synchronization process can recover if there is a disruption during data migration, if transfer and sync does not complete in time cutover cannot be completed) to try various alternatives to resolve data migration issues. During this migration the organization and Deloitte leveraged several different tools and technologies for data migration depending on which tool was the best fit such as AWS Data Migration Service (DMS), Doubletake, SQL server replication, Rsync and traditional backup and restore.
Handling complexities and unexpected issues during production cutovers despite rigorous planning and preparation: The scale of the unexpected COVID-19 pandemic resulted in travel bans, lockdowns, and inability to co-locate. Deloitte managed disruptions due to COVID-19 and executed migrations remotely instead of co-locating through a series of measures such as simulating mock cutovers, securing contingency hardware, deploying multiple collaboration tools, onboarding back-up resources.
Teams spread across geographies: Large global media organization as in this case have teams spread across the world. Complex data center exit projects like this require discovery & analysis, requirements gathering, designing the appropriate solution the target state for migration. Discovery and analysis require a lot of interaction with application owners, subject matter experts, with team spread across multiple geographies introduces several challenges in alignment and coordination. To mitigate these challenges the team planned for longer lead times for activities associated with teams located in different geographies and collocated team members in respective geographical location. Deloitte team focused on comprehensive communication, meticulous planning, and coordination. It is important that the team is asking right questions and that the questions are being properly documented.
Diverse set of applications to be migrated: During the project, constraints imposed by organizational changes resulted in a highly diverse set of applications to be migrated making high-touch migration approach a necessity. ATADATA or CloudEndure approach works best for applications and its infrastructure with little or no shared resources, fully cloud compatible OS and application stack. The Deloitte team wrote custom scripts (complying with the organization’s Cloud COE and AMS guidelines) to perform tasks such as data migration, database deployment, schema generation and infrastructure deployment.
Accomplishments and Value Add
Deloitte, partnering with AWS, successfully managed the migration execution on time on budget and ensured deadlines were not missed, applications cutover smoothly and performance met customer expectations. But Deloitte continuously found opportunities and added value to the program by constantly optimizing and recommending economic ways of execution. Optimization and economizing were a continuous effort while the migration was inflight. Some other accomplishments and Deloitte value added are:
Managed aggressive timelines: Adhered to aggressive data center exit timelines, decommissioning 5000+ servers through meticulous planning and automation. Divided the migration into multiple cohorts – established governance framework to manage multiple teams involved for on-time, on-budget delivery over a 24-month period.
Managed complexities & unplanned circumstances: Executed migrations of complex inter-dependent applications with synchronization, encryption, and failover capabilities for high availability requirements. Leveraged Deloitte Cloud Workbench and Migration Factory approach to create detailed budget and frictionless migration process. Managed COVID-19 disruption and executed migrations remotely instead of co-locating through a series of measures (such as simulating mock cutovers, securing contingency hardware, deploying multiple collaboration tools, onboarding back-up resources, etc.).
Created re-usable & cloud-native solutions: Worked alongside Enterprise Architecture and AMS to address technical issues, creating dozens of new solutions and integration patterns. Designed, developed, and implemented automation scripts for application migration and on-going deployment. Worked with client and AMS to build custom autoscaling solutions for legacy applications to scale during peak season demand.
Coordinated team efforts: Seamlessly managed long working sessions and cutovers requiring 40-60 hours window without impacting deliverables through follow the sun model ensuring detailed handovers between US/USDC and USI counterparts to eliminate knowledge gap and reduce dependency.
AWS Availability Zones (AZ) are discrete data centers with redundant power, networking and connectivity within an AWS region. Redundancy and high availability (HA) were built in the target state using multiple AZs. Even though it might appear to be a straightforward DATA CENTER exit, Deloitte did not just do lift-and-shift, but also re-architected the solution looking from the point of view of high availability. In some cases, HA was maintained as-is (on-premises) in target state (AWS) in other cases better HA in target state (AWS) was achieved.
Cloud migrations can be complex whether driven by data center exit or enterprise application modernization strategy. These migrations present many challenges as discussed in this blog and it requires meticulous planning and surgical execution.
Deloitte assembled a team of Cloud Tech Advisory and Delivery experts to address the high touch needs of the project. The team executed the migration in two phases, Phase-I (Assessment) and Phase-II (Migration execution). Deloitte performed planning, design, implementation, cut over efforts in lock step with client’s product, engineering, security teams, AWS Professional Services, Amazon Managed Services teams and other partners.
Even though the primary objectives were to exit the data centers on schedule and on budget, Deloitte ensured that target state application architecture followed AWS Cloud architectural best practices/principles and pro-actively identified cost optimization opportunities throughout the course of the project. Several optimization levers were exercised to determine the target state architecture. Those key levers included right-sizing, resource tagging, license, database and storage optimization.
During the execution there were many challenges faced by the team that needed to be addressed; challenges such as aggressive timelines, complex application dependencies, wide range of skills and people to be deployed across geographies. Deloitte tailored a custom program plan, built a team and operating model to meet the project needs. Successfully delivered during COVID-19 pandemic period when remote working was the only choice without impacting schedule, quality, and the client’s requirements.
At the end this is a great success story that shows how complex migrations can be executed with a well-designed migration execution plan, proper alignment and partnership between Enterprise, AWS Pro-Serve, and Amazon Managed Services.