Skip to content

Uncategorized

Rules For A Cloud Data Management System (CDMS)

A Cloud Data Management System is a modern superset of a Relational Database Management System, designed for meeting the requirements of 21st century applications and deploying infrastructures. Apart from delivering the complete range of RDBMS capabilities that include ACID transactions, SQL and supporting all the APIs and tools that come with them, this system must:

  • Meet increasing workload demands,
  •  Support advanced datacenter hardware as well as management frameworks,
  •  Support the non-SQL paradigm
  •  Handle unstructured and structured datasets

In particular, a CDMS should embrace the flexible, dynamic and modern cloud computing environments. Following are twelve parameters that a CDMS must meet to perform best in this advanced technological era.

 1.Scale Out For Optimum Performance

A good CDMS is the one that delivers capacity on demand by deleting or adding storage and computational resources to a running database. It should have the elasticity of scaling out to high transaction volumes, i.e., in the millions of TPS (Transaction Per Seconds), and web-scale database sizes, by adding virtual or real machines, storage systems and networks to a live database. Likewise, a CDMS must also be able to scale in gracefully when you do not need resources anymore.

2.Single Logical Database

Irrespective of the complexity of the application, an effective CDMS presents you with the view of a logical, single, always available and consistent database. It shields you from employing explicit partitioning or caching tactics to accomplish massive database scalability. In fact, CDMS must encapsulate or obviate these complexities, so that an administrator or developer can easily use the database regardless of the complexity or scale.

3.Run and Scale On All Infrastructures

A CDMS must have the ability to run on all infrastructures like single machines, public clouds, private clouds and combinations of the aforementioned. It must be capable of running in a heterogeneous environment that can incorporate various virtual machines, network infrastructures or operating systems. A CDMS must excel on enterprise as well as commodity hardware equally.

4.Active Geo-Distribution

To support the geographically distributed workload, disaster recovery and always-running applications, a CDMS must have the ability to run concurrently in numerous datacenters. This means, a CDMS should deliver active operations having consistent semantics, should work between and across (WAN) Wide Area Networks and understand the ways of localizing cache.

5.Non-Stop Availability

An efficient CDMS has the capability of running non-stop for months or even years without failing. A CDMS can never have one point of failure. In fact, it must presume the failure of infrastructure and must be able to detect it automatically, handle the changes in system and recognize extreme events like network partitions. It should be able to take decision about how to respond to network partitions. It must also support upgrades of operating systems, hardware, CDMS versions and changes to schemas as well as other DB administration tasks without affecting or shutting down the availability of CDMS.

6.Embrace Cloud

A CDMS must run in a cloud environment. It should be designed for supporting the cloud-scale performance requirements and at the same time, being supple against the latency challenges and inherent concurrency. It must not only provide TPS and load rate guarantees but also maintain such rates as concurrent load grows. In other words, a CDMS must support the cloud management frameworks and also integrate with advanced cloud stacks.

7.Distributed Security

For a CDMS, it is important to have enterprise class security both at system level as well as database level, which include:

  • Authenticating and accessing the control of machines before they are included as a part of the trusted group,
  •  Authenticating and accessing control of processes of database before they are permitted to participate in any database,
  •  Encrypting all of the communications between machines,
  •  Database-level security for the database users.

 

8.Empowers Developers and Administrators

 The developers and administrators can benefit from CDMS, provided that it meets the following criteria:

  • It supports rapid development and frictionless evolution of applications.
  •  It must be simple to use, without any time-consuming requirement for provisioning of various database servers, or schemas that affect the application development.
  •  Apart from these, it must be integrated with advanced APIs and programming languages, tools of database development and application development frameworks.
  •  A CDMS should enable you to easily redefine and update data as your applications change.
  •  A CDMS should offer one secure administration point for all its resources and databases. It should simplify auditing, process management, profiling and resource allocation.
  •  It must also enable zero-admin, policy-driven services, managing the system as a whole.

 

Be Sociable, Share!
    The following two tabs change content below.

    Sadaf Ajmal