The Challenge
It was difficult to offer regular, on-premise, classes in their school owing to the Covid-19 safety protocols and averting any risk of contracting and transmitting coronavirus. Thus, Learnix wanted to go online and shift to an LMS (Learning Management System) to continue offering online classes amidst the pandemic. However, their on-premise solution, Digitalocean failed when they tried going online.
Why AWS
AWS, without a doubt, offers the most robust, reliable,
and resilient cloud infrastructure. Software is no
longer a single entity and cloud modernization has made
it possible for multiple software to interact, resulting
in phenomenal solutions.
Learnix understood the impact it would make by enabling
online collaboration and helping students continue their
lessons. Hence, they chose to form a collaboration
between a network of like-minded people and brought them
on a single collaborative platform. This shift to AWS
pivoted Learnix’s business operations to a new high and
connected 22000 learners with online resources
(e-content, technology stacks etc) amidst the Covid-19
pandemic to date.
Why Zero&One
Zero&One is one of the leading AWS Certified Partners that offers services globally with a focus on the MENA region. When you choose to work with us, our experts work in tandem with you to curate a custom solution for your business needs, while maintaining AWS standards of tech enablement.
The Solution
The first thing Zero&One experts did was sit with the
key stakeholders and understand the business use case
which included identifying the need of AWS, end-user and
desired outcome. The most robust solution for them was
using Moodle workload. A separate account was created
for Learnix on AWS. The typical Moodle architecture on
AWS provides autoscaling to meet the high and lows of
the demand curve and ensure a smooth transition without
downtime.
We have faced a challenge during migration process,
choosing the best proper machine type that could hold
the load, but we want this to work on autoscaling
purposes taking into consideration the cost, as we know
the load during afternoon, night, holidays and weekends
is very low compared to normal teaching day, for that
reason we worked on short-lived EC2 instances that are
launched automatically by Auto Scaling Groups based on
an auto-scaling policy that utilizes data from
cloudwatch monitoring services.
The transition to the cloud was performed with AWS CDK
(Cloud Development Kit) which was provisioned through
ClouldFormation. After uploading the template file over
CloudFormation, security credentials for Moodle were
generated and stored using Amazon S3 permissions.
MariaDB was prepared with RDS. The resources were
deployed inside a Virtual Private Cloud (VPC), plus
three Amazon Simple Storage Service (Amazon S3) buckets.
Moodle deployment reliability could be further increased
by configuring the application server on Amazon EC2
elastic load balancing and autoscaling. The content
repository comes from EC2 instances and transitions from
Amazon S3 buckets to come up the CloudFront, which is
again configured as per the end-user login. This means
that a C++ student will be able to access files and
content that is specific to the course and a history
research student will be able to access data as per
their subscription. These EC2 instances are launched
automatically by Auto Scaling Groups based on an
auto-scaling policy that utilizes data from CloudWatch
Monitoring Services. Zero&One also implemented custom
autoscaling for RDS that enables new readers in the
morning and disables them at night.
For a robust security architecture, we insist on using
session manager that provide a secure connection to EC2
instances, we use session manager to avoid the using of
inbound ports, ssh, jump box and bastion hosts. User has
the privilege to connect to any running machine using
one click. In addition to that we don't want all IT
users to have access on all instances, by the help of
IAM policies we were able to control which individual
user can use session manager and which server they can
access, to avoid any wrong access and wrong command on a
production server that could lead to a downtime. In
addition, to that we found that using Inventory is
mandatory, for a better visibility about the
architecture. that would help us in troubleshooting
process in case we face any problem. Inventory will
allow us to collect all parameters that are related to
network, application, operating system name and
version,IAM role, DNS, operating system update time and
date. Third feature we have used is the OpsCenter.
OpsCenter help us in reducing mean time to solve
problems affecting our resources, as we are running like
more than 30+ EC2 machines, RDS, load balancer,
memchache, cloudfront. We have integrated it with
cloudwatch to start receiving OpsItems from alarms, in
case of high cpu utilization, failing of an instance,
EBS reaches maximum capacity, etc... based on the report
we can investigate and diagnose the root cause of the
problem.
Results and Benefits
Digital transformation using AWS cloud proved to be a
game-changer for Learnix. The Amazon EC2 auto-scaling
and elastic load balancing managed the user-specific
cloud front irrespective of users geographic location
and removed downtime. ElastiCache was used to scale
seamlessly and store open-source compatible in-memory
data over the cloud to be used by the end-users. The
magnitude of scale with AWS helped them to cater to
22,000 students amidst the pandemic and the number is
going strong. With automated uptime, the time to market
was reduced and gave enough space in time for higher
value application enhancement for the future.
A detailed TCO analysis showed that the costs were
reduced from $17,000 to $7,000, which means a reduction
in monthly expenditure by 40%. The team is equipped to
work on the future scalability and research, thanks to
the reduced costs, and reliability of the migration and
robust infrastructure to carry operations even with
higher yet distributed and managed workload. Learnix is
now focussing on business more than IT infrastructure
operations and maintenance.
