If the answer is DB2, you'll know that the opportunity is there to get good DBA support right off the bat and share the database with more modern software development on various platforms. If the answer is IMS, brace for a bit of a learning curve and know that it will be harder to get good DBA support - and that the job may turn into a bit of a dead end if they won't switch to DB2, or it may be a huge grind to reverse-engineer everything and migrate it to DB2. The upshot here would be this: the first question you should ask about a mainframe software development position is whether the shop uses IMS or DB2 for databases.
#Zcloud zos code#
With DB2, you can also have mainframers writing solid code in COBOL and other developers doing great things with DB2 in other languages (because SQL and relational databases are not tied down to the mainframe environment like IMS is). Database management (DBA) has evolved into its own highly technical environment with its own skill set, and developers working closely with DBAs can leverage their respective strengths and really produce impressive applications. The other really big thing here is that using DB2 (a standard relational database accessed via SQL) lets you split off database management into a separate team from the actual software development team. IMS was developed in the 60s and is still in use in a lot of applications, but it's not a skill set that's popular and it's a bit limited compared to SQL on DB2. I've written large mainframe applications with both of these database systems and have gotten excellent performance every time, but IMS is definitely on the wane whereas DB2 is still relevant. There are two big players in that space - IMS and DB2.
One big thing to look at that you haven't mentioned is which backend database system you use on the mainframe. I'd bet mainframes are here to stay for AT LEAST 30 years. It's just not sexy when you're talking with your friends who dig the cloud mumbo-jumbo. So yeah, if you wanted to jump on the "Master the Mainframe" and learn COBOL, you'll surely have lots of well-paid work. Today, all major banks and airlines rely on z/OS to keep running and they're all short on people to maintain these systems. They worked 7 years on it, and were able to migrate only 20% of the code. How are you going to rewrite that?Ī friend of mine who moved to the UK was working on a project to rewrite HSBC's mainframe code into Java. Most cases when I talk with people who maintain the mainframes, they have no documentation whatsoever for critical routines. Code made by people that may already be dead. The big issue with migrating off from mainframes is that you have to rewrite LOTS of code. But with pervasive encryption on CPU and memory. Z/OS still is the most reliable hardware platform out there. Well, COBOL isn't classy, isn't trendy, isn't "cloudy" but it is damn reliable, dependable and fast.
0 kommentar(er)
