Is it time to consolidate your Java apps? Onto a mainframe?

Is it time to consolidate your Java apps? Onto a mainframe?

Summary: For the better part of three decades now, IBM's mainframes -- often referred to as big iron -- have been coming under assault from smaller iron: minicomputers and servers that may not by themselves have the sheer horsepower to keep up with a mainframe but can often get the job done, especially when a bunch of them work together on the same tasks (a technique known as scaling out).

SHARE:
TOPICS: Open Source
1

For the better part of three decades now, IBM's mainframes -- often referred to as big iron -- have been coming under assault from smaller iron: minicomputers and servers that may not by themselves have the sheer horsepower to keep up with a mainframe but can often get the job done, especially when a bunch of them work together on the same tasks (a technique known as scaling out).  The newer smaller, and always improving technology has kept IBM on its toes in terms of the sorts of work mainframes can do, how mainframes are packaged, and what they cost to use. 

According to James Governor, principal analyst at IT research outfit Redmonk, IBM is once again evolving.  This time, it has repriced it's mainframes for those customers running Java applications.  Wrote Governor,

Java workloads can be consolidated onto zAAP, without paying mainframe fees for the capacity, basically eliminating a major problem for mainframes, which is that the traditional licensing model is based on total system capacity (measured in MIPs). But with IBM's offload processors "modern" workloads- Java, Linux and in some cases relational data - are not counted towards the MIPs total, but are licensed separately.

The message from Big Blue is that with its mainframes, not only are they big enough for you to think about consolidating your Java apps onto one system, the brute force of a mainframe may also make them run faster and, because of the way IBM is pricing its mainframes for Java-based application, it may very well end up being easier on the budget too.  The post caught my eye because of a podcast interview I did last week with executives from Mainsoft and Azul Systems. 

With it's Network Attached Processor for scaling Java apps up, Azul in particular is also focused on the same sort of problem that IBM seems to be going after: consolidating Java apps, making them run faster, and reducing cost.  One difference is that IBM makes Java apps faster with brute force.  Azul does the same but has also tweaked Java under the hood much the same way BEA has tweaked its JRockit Java Virtual Machine to make it more efficient.  Azul is also considering building the same sort of solution for .Net which left me wondering if why companies may be focusing on Java first and then turning their attention to .Net.  Is it a sign that Java has more marketplace traction?  According to Governor, probably not. 

In my podcast interview of him (available as an MP3 that can be downloaded or, if you’re already subscribed to ZDNet’s IT Matters series of audio podcasts, it will show up on your system or MP3 player automatically -- see ZDNet’s podcasts: How to tune in), Governor sees many third parties focused on making Java or .Net more efficient as having picked their battles (unlike Azul which appears to be going for both).   In the interview, we cover IBM's Java mainframe strategy, we talk about the larger .Net vs. Java picture, and Governor predicts that Sun will be making a relatively big open source announcement at JavaOne coming up next week.

Topic: Open Source

Kick off your day with ZDNet's daily email newsletter. It's the freshest tech news and opinion, served hot. Get it.

Talkback

1 comment
Log in or register to join the discussion
  • Being the Last on a TimeShare

    Mainframes bring with them the efficiency of specialization and division of labor. In the PC world the user often is also the operator, systems administrator, backup/restore specialist, network administrator, report designer and programer. These rolls consume the end users time but that consumption was preferable to timeshare on a mainframe when the number of users paying for the operation of the mainframe became smaller and smaller.

    What happens then is that fees increase so that the costs of mainframe operation can be recouped by the mainframe owner.

    We now face the odd situation that Microsoft is faced with fewer and fewer paying users. As that pay base goes down, Microsoft will do the same as the timeshare mainframe owners. Inotherwords the last user on Microsoft software ends up paying for all the costs of supporting that software - just like in the golden age of mainframes.

    There is no silver bullet. Gartner likes to claim that the IT workforce will be reduced owing to outsourcing and efficiencies and the "demystifying" of the craft but forgets that prior to computer science programs at universities all businesses maintained staff called business analysts or system analysts which really is the same thing as IT. In addition there was a lot of nepotism. The bosses son or daughter typically became a system or business analyst and today you see those same folks in IT.

    You do not want to be the last user on a mainframe anymore than you want to be the last user paying Microsoft.

    Frank L. Mighetto CDP
    mighetto