Cloudant provides database-as-a-service — something that some companies may want to consider for management help and cloud-size scaling ability. Nick Hardiman describes how DBaaS works.
Cloudant offer distributed database as a service (DBaaS). They are clever database guys in Boston, USA and Bristol, UK (some of them worked on big data produced by the Large Hadron Collider in CERN - that kind of clever). Cloudant provide a managed service for companies that need NoSQL databases but don't want to manage the systems.
A summary of the modern database world
Here are a few examples of products and providers organized into that favourite of analysts, the quadrant.
|Hosted or Managed|
What do those columns mean?
- A traditional organization with ordinary data storage needs uses a system that talks SQL.
- Modern heavy data users and speed freaks go NoSQL.
- An organization with the resources to get down and dirty with scaling, fail-over and monitoring does DB DIY.
- A company that doesn't have the time to deal with MapReduce, sharding or index tuning passes the management to a service provider.
SQL or NoSQL
Relational database systems that talk SQL have been around for decades and permeate IT. The web is built on the open-source MySQL, corporations use closed-source Oracle, and applications contain embedded SQLite engines. Most projects don't require cloud-size scale-out and do just fine with these traditional SQL offerings.
The new database world includes NoSQL products that lose some relational functionality and gain distribution and scalability. Enterprises that must distribute their data across the globe or manage vast amounts of data each day may choose an open source product like Cassandra or Riak.
Cloudant offer a kind of Apache CouchDB hosting and management service. They can distribute databases across multiple cloud providers in many countries. With the Amazon DynamoDB service, a database can span multiple AZs (Availability Zones) but it can't be distributed globally - for instance, across the US-West, EU and Asia Pacific regions.
DIY or managed
The enterprise database world is traditionally managed on site. Even organizations that are heavy users of cloud services often keep their databases on site, to avoid getting bitten by disk IO. These DB DIY enterprises probably own racks of servers and employ teams of DBAs to make it all work.
Growth may lead to the managed option being cheaper than DIY. If the company moves onto another continent, they may need to duplicate the servers and expertise, and work on replicating the data. The work may be more expensive than handing over the data cluster to a specialist management company. Managed covers quite a bit of ground. A company may take care of only the hosting or a company may take care of everything from tuning to development advice.
For companies that don't want to build up a knowledge of database management, they can leave the hard work to others and rent a hosted service. Many providers offer a managed relational database service.
How Cloudant exploit their skills
As a general rule, Cloudant put the customer's data near the customer and build on what the customer already uses.
Some customers have their data contained within one cloud provider, and some have the database replicated across multiple clouds. Cloudant make an effort to be cloud agnostic — they host databases with Rackspace, Amazon, SoftLayer, Joyent and Microsoft Azure. New customers can choose from half a dozen hosting sites on Cloudant's free sign-up page.
Cloudant takes care of the operational challenges.
- managing the little differences between cloud providers
- building automation technology with chef and opscode, to quickly spin up to meet demand
- distributing data behind the scenes
- keeping everything always on and always available
If a development company has a team of developers writing an application for Apple's iOS, they probably want to concentrate on writing the core application and not on creating the data store. Cloudant can do the work of embedding a CouchDB database in the IOS application, syncing that data with the cloud when the device connects to the network, distributing that data within the cloud, and making it all happen without customer intervention.
Hothead Games from Vancouver created a game portfolio called Big Win Sports. Hothead wrote the apps and databases themselves. They were victims of their own success - the games were profiled on the iOS store, a million people played the games and their databases couldn't scale fast enough. Cloudant now manage a much-expanded set of databases for Hothead.The new division of labor
Cloudant could not exist several years ago, before the IT shifts brought us NoSQL and cloud computing. Customers rely on Cloudant to manage the complexity of modern databases, and Cloudant rely on hosting providers to provide infrastructure. Cloudant are part of the new division of labor in the global economy, where geographical boundaries are increasingly irrelevant.