Majoriy of the typical application design warrants an Application server and database servers. The application servers interact with database server instance using connection/pooled connections. The interaction between app server and database server is highly I/O intensive and performance is good when these two instances(App server and DB) are on the same physical server or atleast better when they are on same LAN.
Is it advisable having to rely on Relational Database Services from Amazon and an appserver running elsewhere?
May be if the application runs on EC2 cloud and access Relational Database Services, it offers required performance. However, if we consider a small application, paying 11cents/hr for an EC2 instance and another 11 cents/hr for an RDS instance may be an expensive affair.
If it is expensive for small applications what are the other scenarios where one can benifit from Amazon's Relational Database as Service?
1) If the application is a Client-Server based? (If the client is accessing DB over a WAN)
2) Requires backup storage for any OLTP applications.
3) For OLAP applications?
Though Amazon and Google are held as competitors in the cloud space, I believe Google's App Engine approach to cloud computing is far better and leaves the competition far behind. The approach itself is different where it offers a application server and a persistance layer with absolutely zero configuration requirements. Google's persistance layer "bigtable" is a problem for "Relational Database" lovers. Amazon, with its huge infracture and R&D team, could offer a simiar package - may be with Tomcat and MySQL, most widely used combination among small application developers. (Lately being adopted at enterprise level applications as well).
Is it advisable having to rely on Relational Database Services from Amazon and an appserver running elsewhere?
May be if the application runs on EC2 cloud and access Relational Database Services, it offers required performance. However, if we consider a small application, paying 11cents/hr for an EC2 instance and another 11 cents/hr for an RDS instance may be an expensive affair.
If it is expensive for small applications what are the other scenarios where one can benifit from Amazon's Relational Database as Service?
1) If the application is a Client-Server based? (If the client is accessing DB over a WAN)
2) Requires backup storage for any OLTP applications.
3) For OLAP applications?
Though Amazon and Google are held as competitors in the cloud space, I believe Google's App Engine approach to cloud computing is far better and leaves the competition far behind. The approach itself is different where it offers a application server and a persistance layer with absolutely zero configuration requirements. Google's persistance layer "bigtable" is a problem for "Relational Database" lovers. Amazon, with its huge infracture and R&D team, could offer a simiar package - may be with Tomcat and MySQL, most widely used combination among small application developers. (Lately being adopted at enterprise level applications as well).
Comments
Post a Comment