For more than 30 years, mainframe computers have been the backbone of computing systems throughout the world. Even today it is estimated that some 80% of the worlds' data is held on such machines. However, new business requirements and pressure from evolving technologies, such as the Internet is pushing these existing systems to their limits and they are reaching breaking point. The Banking and Financial Sectors in particular have been relying on mainframes for the longest time to do their business and as a result it is they that feel these pressures the most. In recent years there have been various solutions for enabling a re-engineering of these legacy systems. It quickly became clear that to completely rewrite them was not possible so various integration strategies emerged. Out of these new integration strategies, the CORBA standard by the Object Management Group emerged as the strongest, providing a standards based solution that enabled the mainframe applications become a peer in a distributed computing environment. However, the requirements did not stop there. The mainframe systems were reliable, secure, scalable and fast, so any integration strategy had to ensure that the new distributed systems did not lose any of these benefits. Various patterns or general solutions to the problem of meeting these requirements have arisen and this research looks at applying some of these patterns to mainframe based CORBA applications. The purpose of this research is to examine some of the issues involved with making mainframebased legacy applications inter-operate with newer Object Oriented Technologies.