b'Data trends JSON data management. If you already have data in mongo that you wish to migrate to AWS directly, this can be a viable option. This also results in a low vendor lock-in, you can basically take your application and plug it with a MongoDB database hosted anywhere without much effort. However, if there are data sovereignty concerns (for example, data must be hosted in the Australian region) DynamoDB is more flexible as it is available in all regions, while DocumentDB, for now at least, is only available in a small range of regions. Figure 3.Vertical vs Horizontal Scaling (from https://www.researchgate.net/figure/Vertical-Vs-Horizontal- Also, DynamoDB scales better, with no scaling_fig3_293799134)linked. Hence when we have more data, we can just add more resources and scale based on our needs. Particularly considering the size of AEM datasets and more favourable billing outcomes with horizontal scaling on cloudthis can present an ideal solution. Also, a lot of cloud NoSQL database offerings come with some desirable SQL featuresfor example, DynamoDB from AWS having ACID-compliant transaction facility.Now that we have picked NoSQL as ourFigure 4.Pricing can differ across platforms depending on need (from https://www.simform.com/blog/database of choice, we need to pickcompute-pricing-comparison-aws-azure-googlecloud/)a cloud vendor and implementation for it. Or maybe we can pick a solutionand we can take full advantage oflimit on table size, while DynamoDB that can be supported on multiplethat. However, this requires a lot morehas an instance size limit (but its still cloud platforms? This process is farengineering effort with set-up andgenerous). So, if vendor lock-in can more painful and will require moremaintenanceparticularly comparedbe fully embraced without any fear, compromise than if we just pick a cloudwith using a managed multi-cloudDynamoDB is the way to go.vendor and run with it. But there aredatabase solution like MongoDB Atlas or some noticeable benefits. Astra DB. But, if the objective is to reduceGoogle Cloud has an interesting range Making a pluggable software solutionengineering overheads, multi-cloudof choices. Whenever a conversation can avoid vendor lock-in. In the case ofmay not be a great idea as it will makeabout NoSQL databases in Google building a processing toolkit that usersthe solution complex, and migratingCloud comes up, Firestore ends to take utilise online, one with large datasetsfrom one cloud platform to another aa front seat. It is widely popularised and long running algorithms, beingchallenge. by mobile application developers, and locked into a single cloud vendor forIn that case, we can look at the native life means losing out on importantofferings from different cloud providers. bargaining power. As the number ofAWS, Google Cloud and Azure are users scale and the cloud bill becomesevaluated as they have majority of the bigger, having the ability to move tomarket share in Australia and overseas another cloud vendor can facilitate(Figure 5), and when people generally shopping around and reduce billstalk about picking up cloud solutions significantly. Cloud providers willthe discussions mainly focus on one of certainly have different specialisations,these three.quality, and pricing models (Figure 4). Depending on usage some may provideFor AWS, DynamoDB will be the top more benefits over the others, benefitsNoSQL candidate. It is fully managed, has that may not become very apparenta generous free tier, and will certainly before users start to scale. cover development without costing much. It is designed to run high-In that case we can deploy our ownperformance applications at any scale NoSQL database in a Kubernetes clusterand even have some characteristics and use a NoSQL database image ofof SQL databases, like being ACID our choice (for example, MongoDB orcomplaint. Another option is AmazonFigure 5.AWS currently dominates the cloud market (from https://www.statista.com/Cassandra). All cloud providers will haveDocumentDB. This is a MongoDBchart/18819/worldwide-market-share-of-leading-options for a flavour of Kubernetes,compatible database, purpose built forcloud-infrastructure-service-providers/)39 PREVIEW FEBRUARY 2022'