Last week, an unprecedented disruption led to a temporary paralysis of the Amazon Web Services in the US East Region. With websites like Quora and Trello down, users and database managers had no way to validate the consistency of their data or transactions. While some chose to shift their operations to an alternate location, there were many that were left high and dry because of the crash of the Simple Storage Service (Amazon’s S3). Turns out, at the time of the crash, Amazon was hosting the AWSome Day conference in Edinburgh, Scotland, using the pedestal to brag about the impeccable performance of their internet services. The S3 crash took down countless websites across the globe who relied upon the storage services from AWS’ S3. The clients who turned to their respective B2B Services found the latter to be equally confused. However, there were some B2B Services that had mechanisms in place to tackle this unforeseen situation.
Already, AWS is the largest provider of web services across the United States and with server locations across the world, there is no dearth of enterprises to vouch for their credibility. This has even led to multiple B2B enterprises to opt for AWS’ S3 and when the latter crashed, there were an exceptional few who tackled this situation through decoupling and the Fail-Safe mechanism to keep the show going for their clients.
At Juggernaut, we opted for a two-fold approach to tackling this problem. Starting with the Fail-Safe mechanism, we also have an architecture in place where EC2 (Elastic Compute Cloud) is used for serving the applications and their respective data on AWS’ S3. However, in case any of these two entities goes offline or faces a disruption, the other one takes over for the smooth conduction of the operations. As we employed this modular architecture, the disruption that lasted for a few hours was helped by the presence of EC2 which took over from S3. Given its limited memory span, EC2 facilitated the operations for the disruption period, and once the S3 was online, the data was automatically moved from EC2 to S3.
As a B2B Service, our objective is to ensure that the services go on without any disruption, even if the web services we are using are affected by a natural or a man-made disaster.
The buck doesn’t stop here, however. As an enterprise, we have opted for the Cross-Region Replicas in our AWS Management Console. For our clients, we do operate a read replica in a region that is different from the master database region. If there is a regional disruption, like there was one last week, we had the option to promote the replica as the master database region, thus eliminating any probability of our client’s website and app being affected. Executing it, we helped our clients keep their operations going as many of the other prominent websites crashed. A number of features on our AWS Management Console helped us fight the disruption, including the option to migrate between regions. The cross region read replicas made it easy for us to migrate applications from the US-East region to a different one. For Juggernaut and our other clients, we immediately moved it to the Singapore region and promoted it as the master database. Given how our clients have a user database all across the globe, it was of paramount importance to keep the show going, and in the end, we successfully did.
Having a modularized system in place where this least dependency on a single server adds to the credibility, consistency, and efficiency of operations for our clients. Alongside, load balancing for the website and app data helps in the case of a natural disaster. Clearly, no web service wants their servers