It's all about the answers!

Ask a question

Is there a way to achieve Jazz "Multisite"?


0
2
Jirong Hu (1.5k8273256) | asked Aug 06 '11, 8:45 a.m.
I have read the following two articles regarding content proxy caching approach. This approach still requires user in a WAN and have access to the Jazz server.

https://jazz.net/library/article/325
https://jazz.net/blog/index.php/2010/10/04/does-rational-team-concert-support-multisite/

Below is my client's comment, but ClearCase Multisite also has to be in a WAN?:
WAN wont work for us here on this project, as the customer has stipulated that no direct access is allowed outside of Canada to their systems.

This means we need to replicate the repository to a location in Toronto, and then from there to our offshore site. And for changes they make, vice versa. I know ClearCase multisite was setup to produce delta files of its VOB, that we ftpd from site to site. It incorporated those back into the offshore VOB, so everyone was relatively up to date (I think we had a 1hr synch frequency).

Does the caching mechanism described provide a similar feature?

6 answers



permanent link
Geoffrey Clemm (29.7k23035) | answered Aug 06 '11, 9:37 a.m.
FORUM ADMINISTRATOR / FORUM MODERATOR / JAZZ DEVELOPER
If they can ftp from one site to another, why can't they http between
the two sites? (http can be made just as secure as ftp).

Cheers,
Geoff

On 8/6/2011 8:53 AM, hujirong wrote:
I have read the following two articles regarding content proxy caching
approach. This approach still requires user in a WAN and have access
to the Jazz server.

https://jazz.net/library/article/325
https://jazz.net/blog/index.php/2010/10/04/does-rational-team-concert-support-multisite/

Below is my client's comment, but ClearCase Multisite also has to be
in a WAN?:
WAN wont work for us here on this project, as the customer has
stipulated that no direct access is allowed outside of Canada to
their systems.

This means we need to replicate the repository to a location in
Toronto, and then from there to our offshore site. And for changes
they make, vice versa. I know ClearCase multisite was setup to
produce delta files of its VOB, that we ftpd from site to site. It
incorporated those back into the offshore VOB, so everyone was
relatively up to date (I think we had a 1hr synch frequency).

Does the caching mechanism described provide a similar
feature?

permanent link
Jirong Hu (1.5k8273256) | answered Aug 10 '11, 9:13 a.m.
If they can ftp from one site to another, why can't they http between the two sites? (http can be made just as secure as ftp).
Cheers,
Geoff


It's confirmed that they have no access to directly ftp as well. Basically it's a rule that oversea site/India can't access our Canada system directly. So we have two questions now:

1. Is it possible to using the cached proxy method to do multiple hops? e.g. from our site, to their/offshore's Toronto data center, and then from there to their offshore site? This way we don't break the rule. Does cached proxy support that level of replication?

1. Do you also know how IBM is planning to synchronize the RTC repository for their offshore development needs?

Thanks
Jirong

permanent link
Michael Valenta (3.6k3) | answered Aug 10 '11, 11:42 a.m.
FORUM MODERATOR / JAZZ DEVELOPER
Jirong,

You seem to be saying that direct contact between the Canadian office and the offshore site is not possible but contact between the Canadian site and a site in Toronto is possible and contact between this Toronto based site and the offshore site is possible as well. Is this a correct interpretation?

If so, it may be possible to use distributed SCM to create a replica of the components that are being worked on offshore at the Toronto site which can then be replicated offshore (i.e. the customer would need to have 3 RTC servers: one at their Canadian site, one in Toronto and one offshore). Any changes made by either side can then be pushed and/or pulled from the shared replica in Toronto as needed.

permanent link
Jirong Hu (1.5k8273256) | answered Aug 10 '11, 1:05 p.m.
Jirong,

You seem to be saying that direct contact between the Canadian office and the offshore site is not possible but contact between the Canadian site and a site in Toronto is possible and contact between this Toronto based site and the offshore site is possible as well. Is this a correct interpretation?

If so, it may be possible to use distributed SCM to create a replica of the components that are being worked on offshore at the Toronto site which can then be replicated offshore (i.e. the customer would need to have 3 RTC servers: one at their Canadian site, one in Toronto and one offshore). Any changes made by either side can then be pushed and/or pulled from the shared replica in Toronto as needed.


Great information, let me look into that.

Look at this: https://jazz.net/library/article/535/
e.g. we have A (Ministry) -- B (Data Center) -- C (Offshore)

1. Simply A push changes to B and from B to C, and from C to B and B to A.
2. Modification happens only in A and C.


Thanks
Jirong

permanent link
Jirong Hu (1.5k8273256) | answered Oct 04 '12, 9:04 p.m.
I am back to this project after 6 months away. The project is using VPN token in the past , but now they have the following concerns.


* No way of ensuring that the person accessing the network is the person authorized (Someone else could access the network after the authorized person logged in)
* Expensive to have managers fly to offshore locations to issue token to authorized persons.
* Issues noted with performance

Distributed SCM will require three RTC servers which is too costly, almost not acceptable. My questions now are:
1. Most of large companies now have developers in India (we out-source to IBM and IBM to India). These off-shore developers are not in our WAN/domain. What's the best way to allow them access to our RTC server? Is IBM working on a solution?
2. Technically is it possible to place a HTTP server in front of our RTC server to proxy the access. All requests coming from India access this front tire web server, then this web server router the request to the internal RTC server? I am not a network guy so just asking.

Thanks
Jirong

permanent link
John Camelon (1.7k14) | answered Oct 12 '12, 2:18 p.m.
JAZZ DEVELOPER
This requirement is captured @ this work item.  It is on our roadmap but not currently available.  Right now, distributed SCM requires connectivity.  We have heard of customers sending virtual images of servers between sites and then ressurecting them on different subnets and playing tricks with host files.  I can not confirm that this works, but I have heard of some successes.  

JohnC
SCM Lead

Your answer


Register or to post your answer.