Virtual Application Appliances

Cloud Database Administration:
Database administration skills are the core of traditional enterprises in the data center. DBAs with skills in replication, backup, recovery and  clustering have been in high demand over the years due to the way the enterprises have taken care all aspects of the database administration.

However, the advent of the Cloud has shifted much of the DBA duties from the Cloud Consumer  (Enterprises) to the Cloud Providers (EC2, Azure and more).

Is the role of DBA irrelevant in a Cloud world? This article discuss five new skills that regular enterprise DBAs should acquire to make their presence as unique as always.

1. Improve On NOSQL Database Usage And Skills
In computing, NoSQL is a term used to designate database management systems that differ from classic relational database management systems in some way. In a cloud managed environment, NOSQL databases are here to stay, and they have their own advantages especially  in processing large amounts of distributed data.

• Google's BigTable
• Amazon SimpleDB
• Windows Azure Storage BLOB Storage

All the above mean that the Non-Relational database usage is here to stay and the current DBAs should identify appropriate usage and maintenance of them.

2. Capacity Planning in Terms of On-Demand Usage Cost
Over the years, DBAs have been very successful in reactive monitoring of storage usages in terms of  GB, TB of disk space and also in predicting future capacity planning. However, as the entire trust of Cloud Computing is about a chargeback model based on storage, it becomes even more important for the DBA to understand the costing models behind the storage and help the enterprise with the best model possible for a given cloud provider.

• For example, SQL Azure Business Edition databases are charged at $99.99 up to 10 GB per month and goes on based on further storage , Data transfers = $0.10 in / $0.15 out / GB - ($0.10 in / $0.20 out / GB in Asia)
• Similar Amazon RDS On Demand DB Instances cost $2.60 price per hour for a extra large DB instance and similarly data transfers are priced separately. Apart from this the concept of Reserved Instances and Backup storage additionally alter the pricing.

With the additional concepts behind Cloud storage and pricing, it is the DBAs who need to have these details behind them and can provide the organizations about accurate information of how the storage and usage costs of the database will look like in a cloud environment.

3. Multi Tenancy & Data Partitioning
Earlier DBAs have big  servers and databases with unlimited storage at their disposal and most times the databases have been exclusively owned by enterprise vertical domains and  very little DBAs worried about the size of a single database and the  tenants who share the database.

• SQL Azure has a 50 GB Limit per database
• For each DB Instance class, AmazonRDS provides you with the ability to select from 5 GB to 1 TB of associated storage capacity for your primary data set. While the storage is unlimited , provisioning storage on demand rather pre allocating , makes it more efficient.

The new limitations in the database storage and the need to store multiple tenants require DBAs to think of different ways of partitioning data across multiple databases, which will go a long way toward an efficient data organization for Cloud.

4. Monitoring
On-demand instances are provisioned based on the  usage of the application and hourly demands. Also  much of the  instances  are virtual server based and the IP Addresses that connect are virtual and may get changed over time. All these require that much more  monitoring capabilities need to be brought in by the DBAs to make sure how the databases behave to the demands and how they can be managed.

• Amazon RDS makes DB Instance metrics available via theAmazon AWS CloudWatch service. You can access your metrics using theAmazon CloudWatch API Toolsor via the API itself.
• MicrosoftSQL Azure Database enables a subset of dynamic management views to diagnose the performance problems, which might be caused by blocked or long-running queries, resource bottlenecks, poor query plans, and so on.

The above concepts clearly proves that the traditional DBA needs to know new tools and  new concepts behind monitoring a Cloud Database.

5. Data In & Data Out  Between Cloud & Data Centers
For very small organizations, it would almost be impossible for an enterprise to run its entire business on the Cloud. So with the Cloud adoption it will still be a hybrid environment of:

• Continued usage of ‘On Premise' data center for mission critical applications
• SaaS usage for certain scenarios
• PaaS or IaaS usage for other scenarios

In this Hybrid situation, it's important for the DBA to have all the options to move data in and out of the cloud seamlessly in an efficient manner.

With SQL Azure Data Sync CTP2  Microsoft  delivering an update to SQL Azure Data Sync to enable synchronization between on-premises SQL Server and SQL Azure, giving customers greater flexibility in building solutions that span on-premises and the cloud. With SQL Azure Data Sync CTP2, customers can:

• Extend enterprise data to the cloud rather than replace it
• Move workloads to the cloud in stages preserving investment in existing infrastructure
• Extend data to remote offices and retails stores
• Enable new scenarios spanning enterprise and cloud

Similarly  Amazon RDS has multiple tools based on MYSQL Engine that supports loading data from  data center applications into Cloud.

Summary
Cloud has abstracted certain traditional DBA needs in an enterprise:

• Installation
• Backup
• Recovery
• Replication
• Stand By Servers
• Fail Over Servers

Cloud has indeed brought in new DBA skill requirements and hence the role of DBA is as important as before in a cloud managed environment. However, now the DBAs can concentrate more on the business needs and can support the ‘Do More With Less' slogan that marks the  Cloud Platform usage in enterprises.

Related Stories

• Amazon Releases Elastic Load Balancing, Auto Scaling, and Amazon CloudWatch
• Gartner Asks, Can The Cloud Save The World?
• Global Distributed Service in the Cloud with F5 And VMware