We are planning to install InfoSphere Data Architect for data modeling activities. I have gone thru the IBM documentation but still have few queries about installation. As already we are planning to install IIS 11.3 in Linux. Can anyone with your experience please guide;
1- Is InfoSphere Data Architect requires Server & Client side installation like DataStage? Or it's only Client side installation?
2- If Server side installation requires, is IIS 11.3 server suit installation program include DataArchitect also as part of server side installation or it requires separate installation altogether?
3- As seen from system requirements, InfoSphere Data Architect supports Linux & Windows and it looks like single installation program. Hence doubting if we install on Linux how do we connect from Client / windows machine?
Kindly let me know if you need any details in clearing above doubts. Thanks in advance.
Infosphere Data Architect Installation
Moderators: chulett, rschirm, roy
Note that IS 11.3.x requires 64-bit Windows for the Client tier. The client will not install on 32-bit Windows.
As far as running the IS server tiers on Linux, a lot of people do. The Planning, Installation, and Configuration Guide has the information you need.
I think the IDA software is a separate discussion topic from IS. The IDA installs we have done have been on Windows clients only with no server components and no IS clients (our Data Modelers are separate people from DataStage Developers). It may also support an optional IDA server component; I'm not sure, but I don't think is has any dependency on IS.
As far as running the IS server tiers on Linux, a lot of people do. The Planning, Installation, and Configuration Guide has the information you need.
I think the IDA software is a separate discussion topic from IS. The IDA installs we have done have been on Windows clients only with no server components and no IS clients (our Data Modelers are separate people from DataStage Developers). It may also support an optional IDA server component; I'm not sure, but I don't think is has any dependency on IS.
Choose a job you love, and you will never have to work a day in your life. - Confucius
Thank you for your inputs. Actually we thought of having IDA to be installed on Linux where we are going to install IIS i.e. Linux Server so that to have connecting database ports to be opened to IIS server only. But as per your input to the limited usage of IDA by only Data Modelers, we can also opt to install on only client machine i.e. on 64 bit Windows.
And extending to the above query about IIS installation topology where I can't get any input from IBM guide.
Option 1: Desired Topology: Oracle database for metadata rep on Physical server, services tier & engine tier on Virtual machine.
Option 2: Basic Topology: DB2 as default metadata rep + services tier + engine tier on Virtual machine
For Option 1, both servers to be in same sub net is recommended. As we are going to use almost all tool set from suite (DataStage, QualityStage, InformationAnalyzer, InformationGovernanceCatalog - Business Glossary, Metadata Workbench..)
Please advise from your experience if we go with Option 1 do you see any performance issues or do you suggest Option 2 is best and what could be the known reasons?
And extending to the above query about IIS installation topology where I can't get any input from IBM guide.
Option 1: Desired Topology: Oracle database for metadata rep on Physical server, services tier & engine tier on Virtual machine.
Option 2: Basic Topology: DB2 as default metadata rep + services tier + engine tier on Virtual machine
For Option 1, both servers to be in same sub net is recommended. As we are going to use almost all tool set from suite (DataStage, QualityStage, InformationAnalyzer, InformationGovernanceCatalog - Business Glossary, Metadata Workbench..)
Please advise from your experience if we go with Option 1 do you see any performance issues or do you suggest Option 2 is best and what could be the known reasons?
You may also benefit from this IBM Redbook:
IBM InfoSphere Information Server Deployment Architectures
Personally I'm not a big fan of virtual machines unless maybe you can trust they are fully setup and well supported. I will leave it to others to comment further.
IBM InfoSphere Information Server Deployment Architectures
Personally I'm not a big fan of virtual machines unless maybe you can trust they are fully setup and well supported. I will leave it to others to comment further.
Choose a job you love, and you will never have to work a day in your life. - Confucius
-
priyadarshikunal
- Premium Member
- Posts: 1735
- Joined: Thu Mar 01, 2007 5:44 am
- Location: Troy, MI
Virtual machines do provide hardware independence and partitioning of resources for effective utilization but at a penalty of around 10% overhead. For CPU/IO intensive products such as Information server and databases, if VM needs to be used, I ask them to add 20% extra resources (hardware). You may be enlightened to see that Intel did a benchmarking on physical server and on VMware and the results are well documented.
One comment in that benchmark is
The configuration with single VM on the physical server had better throughput. And the graph shows that the throughput decreases with the number of VMs running on the same server.
(unable to copy text, visit page 5-6 for details)
So the benefit of VM, IMO, is easy backup, restore and hardware independence at the cost of 10% hardware resources. Please note that its not VMware player, its VMware vSphere 5.0 which installs directly on the server without OS and supposed to have better hardware control and management with least overhead.
One comment in that benchmark is
The configuration with single VM on the physical server had better throughput. And the graph shows that the throughput decreases with the number of VMs running on the same server.
(unable to copy text, visit page 5-6 for details)
So the benefit of VM, IMO, is easy backup, restore and hardware independence at the cost of 10% hardware resources. Please note that its not VMware player, its VMware vSphere 5.0 which installs directly on the server without OS and supposed to have better hardware control and management with least overhead.
Priyadarshi Kunal
Genius may have its limitations, but stupidity is not thus handicapped.
Genius may have its limitations, but stupidity is not thus handicapped.