The challenge of managing distributed data can only be solved with NoSQL solutions
Managing data over large geographic areas where latency and reliability problems exist is a major challenge for relational databases, but fortunately, NoSQL databases are solving the problem of managing distributed data.
The world of mobile apps, online retailers and social platforms dominate much of the Internet and our online lives. Every day we sign up for a service or create a profile that stores a digital existence in some unseen space. Our expectations of technology have steadily grown to demand quick responses in the collection and retrieval of our data, readily available 24/7. The companies that host these services require versatile environments that are scalable and cost effective to handle this surge of new users manipulating data in unpredictable ways and with third-party vendors. This world didn’t exist 10 years ago and signs show no slowing down. Enter NoSQL.
What is NoSQL?
Our expectations of technology have steadily grown to demand quick responses in the collection and retrieval of our data.Asha Rama
Not only SQL or NoSQL is defined by as being non-relational, distributed, open-source and horizontally scalable. Relational databases, which have been the standard for decades, require strict and rigid schemas that must define types of data before they are entered into tables for later reference. Relational databases, or RDBMS, input data into rows and columns of a table and then relate those entries with other tables. Prior to reading that data the tables must perform a collection or query, and the same goes for writing to these tables. Changing a type of data or modifying tables can cause corruption issues not to mention the time and processing power it takes for the collection of mass amounts of data to occur. NoSQL can be structured or schema-less in collecting and storing data. One JSON document can have multiple attributes that would have spanned multiple tables before. The data's type can be changed without causing relational corruption and any type of data can be entered into a NoSQL database.
What drives NoSQL?
The three main components that drive the use and popularity of NoSQL are big data, big users and cloud hosted databases. Those mobile applications and social platforms that offer a host of features collect vast amounts of information constantly. Droves of various types of data require database resources that can store this content in a quick and easily accessible means. A single mobile app could go viral overnight and suddenly countless users could be registering themselves, each expecting a responsive experience that will keep up with their mouse clicks and finger taps. NoSQL can react to the influx of traffic and users by expanding upward or outward. Instead of overloading expensive and complex database servers requiring high amounts of CPU and memory to support the IO demands, a NoSQL powered environment can distribute its workload among a cluster of virtual or physical machines. Since NoSQL was designed to distribute its operations across member database systems, a server cluster can maintain a quick response time while supporting higher volumes of users. While conducting benchmark’s using the NoSQL solution MongoDB, Intuit found their write times to be 2.5 times faster than using MySQL.
Aspiring startup companies and developers looking for NoSQL solutions have options available, though it would benefit them to understand what their dataset requirements are. There are 4 main datamodels to choose from and many flavors of databases from those types. Some of those models and their corresponding providers are:
| Type |
Solution |
| Key / Value |
Redis, MemcacheDB, etc |
| Column |
Cassandra, HBase, etc. |
| Document |
MongoDB, Couchbase, etc |
| Graph |
OrientDB, Neo4J, etc. |
A brief explanation of each type of data model can be found on Planet Cassandra’s website. Additionally, a full list of 150 NoSQL databases can be found on nosql-database.org's website.
Though NoSQL provides robust and cost effective solutions for invigorated developers and players of the mobile/web app market, there are still some hurdles to consider prior to adoption. Inherently due to its relative newness and mostly open source platform, larger organizations may be hesitant to place themselves on the bleeding edge of technology. Their concerns, which are valid, may include the levels of support and expertise throughout the database community. This isn’t to say that NoSQL providers are stopped in their tracks. InfoWorld, ZDNet and PCWorld each published articles within this last week reporting efforts to curb these concerns for PostgreSQL, EnterpriseDB and MongoDB, respectively. Also, Amazon Web Services (AWS) have competitive options for entirely hosted database services in DynamoDB and a scaled down version in SimpleDB. The video provided for DynamoDB is definitely worth a view for anyone interested in learning more about what hosted services look and feel like. Given the organic evolution the Internet has gone with mobility and web applications, it only makes sense that the foundation of our information evolve as well.
How is NoSQL helping to solve your distributed data problems? Let us know.
