With ever-increasing and varying volume data storage needs posing a challenge for the Architects of Enterprise Platforms everyday now. The solution is in reach. One of the results of amalgamation of many years of experience gained by Alibaba Cloud and resulting varied technologies in Alibaba Cloud is – Table Store. Let’s see what exactly it is.
Simply put it is a NoSQL Database service from Alibaba Cloud. It lets you organize data into tables with seamless scaling utilizing data partitioning and load balancing.
What do we get as a result of this NoSQL database service on Cloud. Well, to begin with:
- Protection of applications from underlying hardware failure.
- SSD based backups. [That is right, we have hard time getting SSDs for production data J]
- High Data Reliability and High service availability.
- Dynamic Through put scaling.
- Unlimited Data Capacity
- Failure detection
- Easy Data Management
- Protection from Network Failures
- Flexible Table Model
- Real Time Monitoring
Let’s read a little bit more about few of the above points.
Dynamic Throughput Scaling: it is important to understand that cloud resources can cost you dearly if not managed well. This is true not just in case of cost – but in performance as well. So for the tables we have the capability to assign the resource of read/write throughput. This is configurable when creating tables. Moreover, this is also doable dynamically.
Easy Data Management: data management could be a humongous task given the complex activities like Data Partitioning, Software/HW upgrades and Cluster resizing – luckily this is all taken care of by TS service itself. So, you can breathe easy and focus on the items that you really need to.
Flexible Table Model: Personally, I don’t remember a time when I could remember all rules defined by Codd for Database RDBMS. Well if you feel that we are in the same boat – then you can still breathe easy as Alibaba Table Store allows different column count for each row.
Not even this – it even allows us to have same columns for different rows to have different data types. Yes, I love that. Devices today are not exactly RDBMS friendly when throwing back data at us.
Real Time Monitoring – is strong and robust enough to get you the number of requests per second and even the average response latency. That could be really helpful in writing APIs that would further finetune the through put in real time.
In next follow up article we will study some use cases where this service would make a helpful impact and ease the development.