sharing an article on InterCloud - Application scaling across multiple Clouds

35 views
Skip to first unread message

Rajkumar Buyya

unread,
Mar 20, 2010, 7:43:30 AM3/20/10
to cloud-c...@googlegroups.com
Dear All:

Australian Research Council (ARC) has awarded a 3-year grant to our
CLOUDS Lab @ Melbourne University, Australia to conduct research on
InterCloud supporting application scaling across multiple Clouds.

We recently wrote a vision paper on InterCloud, which identifies the
need for applications scaling across multiple vendor Clouds,
requirements and research challenges, architectural elements,
"preliminary" performance results, and future directions. This paper
will be published as a keynote paper in upcoming conference (ICA3PP 2010
to be held in South Korea). Preprint version can be downloaded from:
http://www.buyya.com/papers/InterCloud2010.pdf

I hope you will find the paper interesting.

Best regards
Raj

-------------------------------------------------------------------------
Title - InterCloud: Utility-Oriented Federation of Cloud Computing
Environments for Scaling of Application Services

Authors: Rajkumar Buyya, Rajiv Ranjan, Rodrigo N. Calheiros

Abstract: Cloud computing providers have setup several data centers at
different geographical locations over the Internet in order to optimally
serve needs of their customers around the world. However, existing
systems do not support mechanisms and policies for dynamically
coordinating load distribution among different Cloud-based data centers
in order to determine optimal location for hosting application services
to achieve reasonable QoS levels. Further, the Cloud computing providers
are unable to predict geographic distribution of users consuming their
services, hence the load coordination must happen automatically, and
distribution of services must change in response to changes in the load.
To counter this problem, we advocate creation of federated Cloud
computing environment (InterCloud) that facilitates just-in-time,
opportunistic, and scalable provisioning of application services,
consistently achieving QoS targets under variable workload, resource and
network conditions. The overall goal is to create a computing
environment that supports dynamic expansion or contraction of
capabilities (VMs, services, storage, and database) for handling sudden
variations in service demands.

This paper presents vision, challenges, and architectural elements of
InterCloud for utility-oriented federation of Cloud computing
environments. The proposed InterCloud environment supports scaling of
applications across multiple vendor clouds. We have validated our
approach by conducting a set of rigorous performance evaluation study
using the CloudSim toolkit. The results demonstrate that federated Cloud
computing model has immense potential as it offers significant
performance gains as regards to response time and cost saving under
dynamic workload scenarios.

Comments: 20 pages, 4 figures, 3 tables, conference paper
ACM-class: C.2.4

Reference/Cite As: Rajkumar Buyya, Rajiv Ranjan, Rodrigo N. Calheiros,
InterCloud: Utility-Oriented Federation of Cloud Computing Environments
for Scaling of Application Services, Proceedings of the 10th
International Conference on Algorithms and Architectures for Parallel
Processing (ICA3PP 2010, Springer, Germany), Busan, South Korea, May
21-23, 2010.
---------------------------------------------------------------------------


Thiago Baldim

unread,
Mar 20, 2010, 1:24:13 PM3/20/10
to cloud-c...@googlegroups.com
That's is a nice paper. Congrats Buyya

2010/3/20 Rajkumar Buyya <r...@csse.unimelb.edu.au>
--
~~~~~
Register Today for Cloud Slam 2010 at official website - http://cloudslam10.com
Posting guidelines: http://groups.google.ca/group/cloud-computing/web/frequently-asked-questions
Follow us on Twitter http://twitter.com/cloudcomp_group or @cloudcomp_group
Post Job/Resume at http://cloudjobs.net
Buy 88 conference sessions and panels on cloud computing on DVD at http://www.amazon.com/gp/product/B002H07SEC, http://www.amazon.com/gp/product/B002H0IW1U or get instant access to downloadable versions at http://cloudslam09.com/content/registration-5.html

~~~~~
You received this message because you are subscribed to the Google Groups "Cloud Computing" group.
To post to this group, send email to cloud-c...@googlegroups.com
To unsubscribe from this group, send email to cloud-computi...@googlegroups.com

To unsubscribe from this group, send email to cloud-computing+unsubscribegooglegroups.com or reply to this email with the words "REMOVE ME" as the subject.



--
Thiago Baldim

Miha Ahronovitz

unread,
Mar 20, 2010, 4:50:20 PM3/20/10
to cloud-c...@googlegroups.com
Rajkumar,

You are making history. I just discovered your company Manjrasoft Aneka
developement platform and it is impressive, at least reding it on the
Web pages.
http://www.manjrasoft.com/

For other on list, please see Table 1 in Section 2 of the paper from
Rajkumar

Cheers,

miha

mij123.vcf

Ray Nugent

unread,
Mar 21, 2010, 2:35:35 AM3/21/10
to cloud-c...@googlegroups.com
This is old news. These folks are not making history, rather they are bolting old technology to new use cases. There is no innovation here. The funny part is these guys suggesting compute resources should be subject to a commodities market. Sure, let's invite Goldman Sachs to manipulate the datacenter like they did oil, gold, food etc via their HFT platform...


From: Miha Ahronovitz <mij...@sbcglobal.net>
To: cloud-c...@googlegroups.com
Sent: Sat, March 20, 2010 1:50:20 PM
Subject: Re: [ Cloud Computing ] sharing an article on InterCloud - Application scaling across multiple Clouds

Zul Kagalwalla

unread,
Mar 21, 2010, 10:58:58 AM3/21/10
to cloud-c...@googlegroups.com
Raj,

Great paper. You do touch on security as a concern by cloud consumers.

Where in your architecture you provide assurance about security. Probably
SLA parameters will have it.

But still someone needs to keep the security audit level/compliance of the
cloud provider and even the cloud consumer. I am sure the malware guys will
figure out a way to inject bots, virus etc into a VM at some stage. So they
cloud consumer could be a security hacker/attacker.

At some point I expect or its most probably happening, the cloud consumer
will make sure the data they are sending is virus free and data that enters
after being served is virus free. Same applies to the cloud provider. But
still both consumer and service provider would want to make sure they are
dealing with a secure provider/consumer.

Also the cloud provider may want to know in SLA or other parameter what the
expected outcome of the service should be i.e. maybe it crashes the VM or
may crash it, since it is a development effort or a QA service.

-----Original Message-----
From: cloud-c...@googlegroups.com
[mailto:cloud-c...@googlegroups.com] On Behalf Of Rajkumar Buyya
Sent: Saturday, March 20, 2010 7:44 AM
To: cloud-c...@googlegroups.com
Subject: [ Cloud Computing ] sharing an article on InterCloud - Application
scaling across multiple Clouds

Dear All:

Best regards
Raj

Tarry Singh

unread,
Mar 21, 2010, 1:05:40 PM3/21/10
to cloud-c...@googlegroups.com
Sorry Raj paper is fine but your company you should be well past writing vision papers and get over the academic hangovers with the university you work for. I agree with Ray. It's just bolting old stuff (grid et al) into new one. 

On another note, can you tell us something more interesting as in:
- Do you have any customers that are running your Aneka stuff in test/production?
- Are you making any progress by introducing new stuff like security, billing etc that others are conveniently avoiding?

Tarry

shalooshalini

unread,
Mar 22, 2010, 1:06:27 AM3/22/10
to Cloud Computing
Raj,

>>
The overall goal is to create a computing
environment that supports dynamic expansion or contraction of
capabilities (VMs, services, storage, and database) for handling
sudden
variations in service demands.
>>

VMotion, Zenmotion features are based on similar premise - but in
actual - in live data centers - how many really use such features?
Why would someone spend money doing this across clouds? What is the
business case and ROI?

Shaloo

Srini Chari

unread,
Mar 22, 2010, 7:35:16 AM3/22/10
to cloud-c...@googlegroups.com

I think there is space for all contributions in the “cloud”. The paper is good. The fact that people are contributing is by itself good.

Ray Nugent

unread,
Mar 22, 2010, 2:25:08 PM3/22/10
to cloud-c...@googlegroups.com
Great point. The use case for moving a live, running stack is pretty much a corner case and it's problematic if the stack is running an app that needs large data to move along with it. There are better ways to achieve similar results of relocating workloads across clouds and, so far, there are not a lot of financial or business drivers either.

But my earlier point is that this paper is just CORBA for clouds. Lot's of prior art on this subject.

Ray


From: shalooshalini <shaloo....@gmail.com>
To: Cloud Computing <cloud-c...@googlegroups.com>
Sent: Sun, March 21, 2010 10:06:27 PM
Subject: [ Cloud Computing ] Re: sharing an article on InterCloud - Application scaling across multiple Clouds

Ray Nugent

unread,
Mar 22, 2010, 2:29:24 PM3/22/10
to cloud-c...@googlegroups.com
I agree with that basic premise. Contribution is good. The original claim was that it was a "Vision" paper. I guess that was my hot button issue as it's not visionary. Rather, it's the application of existing concepts and constructs to clouds. If it was a vendor doing that we would call it Cloudwashing...


From: Srini Chari <ch...@vetrei.com>
To: cloud-c...@googlegroups.com
Sent: Mon, March 22, 2010 4:35:16 AM
Subject: RE: [ Cloud Computing ] sharing an article on InterCloud - Application scaling across multiple Clouds

Tarry

to be held in South Korea ). Preprint version can be downloaded from:

Processing (ICA3PP 2010, Springer, Germany ), Busan , South Korea , May

gopu.r...@wipro.com

unread,
Mar 23, 2010, 6:28:31 AM3/23/10
to cloud-c...@googlegroups.com

Ray

 

With this Technology ,

1.Moving Live machines results in freeing resource/optimizations  is obviously good for better performance on shared environment . As the availability of resource is guaranteed for priority VM’s .

2.On Contrary VM’s access on shared network bandwidth will results to poor connectivity access. With this technology, moving VM’s across non utilized Cloud will efficiently manage the Load. As per demand based  on closest Geo - Regions either for Cloud Hosting providers /On premises Cloud.

3. The time taken to do effort building new Infra on the cloud and migrating the workload and load balancing with APP always take some amount of time.

 

Work Load Migration will be intangible cost savings towards ROI. As there is zero downtime on workload migration as committed by technology VMotion, Zenmotion . This features are basic essential for high availability and Resource Pooling . The real value is realized  as per application Infrastructure SLA penned and Business Value /ROI is obtained as per the SLA.

 

 

 

Warm Regards

 

R.Gopu

 

From: cloud-c...@googlegroups.com [mailto:cloud-c...@googlegroups.com] On Behalf Of Ray Nugent


Sent: Monday, March 22, 2010 11:55 PM
To: cloud-c...@googlegroups.com

Jan Klincewicz

unread,
Mar 23, 2010, 11:35:00 AM3/23/10
to cloud-c...@googlegroups.com
Live migration is most often used to avoid downtime (such as doing hardware maintenance on hosts.)  Actually, Vmotion and XenMotion are both limited to a large extent by heterogeneous servers.  The CPU models (and even steppings) need to be quite similar to achieve this.  This is mitigated somewhat on the AMD Opteron platforms,  however because of the market share enjoyed by Intel, there are far more cases where it is impractical than not.

One can relax the compatibility criteria manually, but at risk of the migration failing.

Dynamic Resource Scheduling (load balancing based on the compute resources needs of individual VMs) is very helpful in getting the best perfromance out of a host, but again, doing so with minimum downtime can be quite risky, and the risk multiplies as the server farms grow and age.

How many Cloud Providers have a homogeneous enough universe of servers to accomplish this in any meaningful scale ?
Cheers,
Jan

Mohamed El-Refaey

unread,
Mar 23, 2010, 11:44:15 AM3/23/10
to cloud-c...@googlegroups.com
Definitely, you are right Ray.
IMHO, Having this capability on the hypervisor level (Case XenMotion and Vmware VMotion) or on the inter-cloud level, is invaluable for data center and infrastructure services' providers. it might be not yet in real use in production, but it is not a big deal to get to use it in true real cases; and the investment in tools and techniques to adopt and improve this capability is worth the time and effort.

Regards,
Mohamed

Rajkumar Buyya

unread,
Mar 24, 2010, 4:12:32 AM3/24/10
to cloud-c...@googlegroups.com

Thiago Baldim wrote:
> That's is a nice paper. Congrats Buyya

Thank you, Thiago.

Cheers
Raj
>
> 2010/3/20 Rajkumar Buyya <r...@csse.unimelb.edu.au
> <mailto:r...@csse.unimelb.edu.au>>

> <mailto:cloud-c...@googlegroups.com>


> To unsubscribe from this group, send email to
> cloud-computi...@googlegroups.com

> <mailto:cloud-computi...@googlegroups.com>


>
> To unsubscribe from this group, send email to
> cloud-computing+unsubscribegooglegroups.com

> <http://unsubscribegooglegroups.com> or reply to this email with the

--
------------------------------------------------------------------------
Professor Rajkumar Buyya
Director, Cloud Computing and Distributed Systems (CLOUDS) Lab
Dept. of Computer Science and Software Engineering
The University of Melbourne
ICT Building, 111, Barry Street, Carlton
Melbourne, VIC 3053, Australia
Phone: +61-3-8344 1344 (office); +61-0421 813 723 (home)
Fax: +61-3-9348 1184; Email: r...@csse.unimelb.edu.au
URL: http://www.buyya.com | http://www.gridbus.org/~raj
------------------------------------------------------------------------

Rajkumar Buyya

unread,
Mar 29, 2010, 1:16:00 AM3/29/10
to cloud-c...@googlegroups.com
Hi Tarry:

Yes, I am still balance research and practice - looking forward to put
more efforts on Manjrasoft side as we progress.

Regarding you question:

> - Do you have any customers that are running your Aneka stuff in
> test/production?

A number of users using our Aneka software are using it for many real
work applications. For example, one of our customers in China (GoFront
Group, a division of Southern China Railway) has utilised Aneka for
high-speed rendering of engineering design frames. In this case, they
build a private Cloud by harness existing IT assets (e.g., a collection
of LAN connected desktop computers) to speed up computations. Please
check out a case study:
http://www.manjrasoft.com/manjrasoft_rendering_casestudy.html
Others are running their applications developed using Aneka and
executing on public Clouds like EC2 (without any change). Check out our
paper on High Performance Cloud Computing:
http://www.buyya.com/papers/HPCC-ISPAN2009-Keynote.pdf

I hope you will more these interesting (to start with).

Cheers
Raj

> <mailto:cloud-c...@googlegroups.com>


> To unsubscribe from this group, send email to
> cloud-computi...@googlegroups.com

> <mailto:cloud-computi...@googlegroups.com>


>
> To unsubscribe from this group, send email to
> cloud-computing+unsubscribegooglegroups.com

> <http://unsubscribegooglegroups.com> or reply to this email with the


> words "REMOVE ME" as the subject.
>
> --
> ~~~~~
> Register Today for Cloud Slam 2010 at official website -
> http://cloudslam10.com
> Posting guidelines:
> http://groups.google.ca/group/cloud-computing/web/frequently-asked-questions
> Follow us on Twitter http://twitter.com/cloudcomp_group or
> @cloudcomp_group
> Post Job/Resume at http://cloudjobs.net
> Buy 88 conference sessions and panels on cloud computing on DVD at
> http://www.amazon.com/gp/product/B002H07SEC,
> http://www.amazon.com/gp/product/B002H0IW1U or get instant access to
> downloadable versions at
> http://cloudslam09.com/content/registration-5.html
>
> ~~~~~
> You received this message because you are subscribed to the Google
> Groups "Cloud Computing" group.
> To post to this group, send email to
> cloud-c...@googlegroups.com

> <mailto:cloud-c...@googlegroups.com>


> To unsubscribe from this group, send email to
> cloud-computi...@googlegroups.com

> <mailto:cloud-computi...@googlegroups.com>


>
> To unsubscribe from this group, send email to
> cloud-computing+unsubscribegooglegroups.com

> <http://unsubscribegooglegroups.com> or reply to this email with the


> words "REMOVE ME" as the subject.
>
>
> --
> ~~~~~
> Register Today for Cloud Slam 2010 at official website -
> http://cloudslam10.com
> Posting guidelines:
> http://groups.google.ca/group/cloud-computing/web/frequently-asked-questions
> Follow us on Twitter http://twitter.com/cloudcomp_group or @cloudcomp_group
> Post Job/Resume at http://cloudjobs.net
> Buy 88 conference sessions and panels on cloud computing on DVD at
> http://www.amazon.com/gp/product/B002H07SEC,
> http://www.amazon.com/gp/product/B002H0IW1U or get instant access to
> downloadable versions at http://cloudslam09.com/content/registration-5.html
>
> ~~~~~
> You received this message because you are subscribed to the Google
> Groups "Cloud Computing" group.
> To post to this group, send email to cloud-c...@googlegroups.com
> To unsubscribe from this group, send email to
> cloud-computi...@googlegroups.com
>
> To unsubscribe from this group, send email to
> cloud-computing+unsubscribegooglegroups.com or reply to this email with
> the words "REMOVE ME" as the subject.

--

Ray Nugent

unread,
Mar 31, 2010, 9:46:35 PM3/31/10
to cloud-c...@googlegroups.com
I agree that motion can be useful. We did a good deal of research on this, filled some patents even. But we concluded that large scale motion of big, running machines would be so chaotic that IT would shun the practice. There are better ways to achieve the same end goal - a goal we all agree on.
Ray Nugent
www.smartscalesystems.com

Greg Pfister

unread,
Apr 1, 2010, 9:12:34 PM4/1/10
to Cloud Computing
Not to dispute your point, but could you say how you achieve the same
end goal or goals without live migration? For example, avoiding
downtime due to scheduled hardware maintenance?

Greg Pfister
http://perilsofparallel.blogspot.com/

On Mar 31, 7:46 pm, Ray Nugent <ranug...@gmail.com> wrote:
> I agree that motion can be useful. We did a good deal of research on this,
> filled some patents even. But we concluded that large scale motion of big,
> running machines would be so chaotic that IT would shun the practice. There
> are better ways to achieve the same end goal - a goal we all agree on.
>
> On Tue, Mar 23, 2010 at 8:44 AM, Mohamed El-Refaey <
>
>
>
> mohamed.elref...@googlemail.com> wrote:
> > Definitely, you are right Ray.
> > IMHO, Having this capability on the hypervisor level (Case XenMotion and
> > Vmware VMotion) or on the inter-cloud level, is invaluable for data center
> > and infrastructure services' providers. it might be not yet in real use in
> > production, but it is not a big deal to get to use it in true real cases;
> > and the investment in tools and techniques to adopt and improve this
> > capability is worth the time and effort.
>
> > Regards,
> > Mohamed
>

> > On Tue, Mar 23, 2010 at 12:28 PM, <gopu.raja...@wipro.com> wrote:
>
> >>  Ray
>
> >> With this Technology ,
>
> >> 1.Moving Live machines results in freeing resource/optimizations  is
> >> obviously good for better performance on shared environment . As the
> >> availability of resource is guaranteed for priority VM’s .
>
> >> 2.On Contrary VM’s access on shared network bandwidth will results to poor
> >> connectivity access. With this technology, moving VM’s across non utilized
> >> Cloud will efficiently manage the Load. As per demand based  on closest Geo
> >> - Regions either for Cloud Hosting providers /On premises Cloud.
>
> >> 3. The time taken to do effort building new Infra on the cloud and
> >> migrating the workload and load balancing with APP always take some amount
> >> of time.
>
> >> Work Load Migration will be intangible cost savings towards ROI. As there
> >> is zero downtime on workload migration as committed by technology VMotion,
> >> Zenmotion . This features are basic essential for high availability and
> >> Resource Pooling . The real value is realized  as per application
> >> Infrastructure SLA penned and Business Value /ROI is obtained as per the
> >> SLA.
>
> >> Warm Regards
>
> >> R.Gopu
>

> >> *From:* cloud-c...@googlegroups.com [mailto:
> >> cloud-c...@googlegroups.com] *On Behalf Of *Ray Nugent
> >> *Sent:* Monday, March 22, 2010 11:55 PM
>
> >> *To:* cloud-c...@googlegroups.com
> >> *Subject:* Re: [ Cloud Computing ] Re: sharing an article on InterCloud -


> >> Application scaling across multiple Clouds
>
> >> Great point. The use case for moving a live, running stack is pretty much
> >> a corner case and it's problematic if the stack is running an app that needs
> >> large data to move along with it. There are better ways to achieve similar
> >> results of relocating workloads across clouds and, so far, there are not a
> >> lot of financial or business drivers either.
>
> >> But my earlier point is that this paper is just CORBA for clouds. Lot's of
> >> prior art on this subject.
>
> >> Ray
>

> >>  ------------------------------
>
> >> *From:* shalooshalini <shaloo.shal...@gmail.com>
> >> *To:* Cloud Computing <cloud-c...@googlegroups.com>
> >> *Sent:* Sun, March 21, 2010 10:06:27 PM
> >> *Subject:* [ Cloud Computing ] Re: sharing an article on InterCloud -

> >>http://groups.google.ca/group/cloud-computing/web/frequently-asked-qu...
> >> Follow us on Twitterhttp://twitter.com/cloudcomp_groupor
> >> @cloudcomp_group
> >> Post Job/Resume athttp://cloudjobs.net


> >> Buy 88 conference sessions and panels on cloud computing on DVD at
> >>http://www.amazon.com/gp/product/B002H07SEC,

> >>http://www.amazon.com/gp/product/B002H0IW1Uor get instant access to


> >> downloadable versions at
> >>http://cloudslam09.com/content/registration-5.html
>
> >> ~~~~~
> >> You received this message because you are subscribed to the Google Groups
> >> "Cloud Computing" group.
> >> To post to this group, send email to cloud-c...@googlegroups.com
> >> To unsubscribe from this group, send email to
> >> cloud-computi...@googlegroups.com
>
> >> To unsubscribe from this group, send email to cloud-computing+
> >> unsubscribegooglegroups.com or reply to this email with the words "REMOVE
> >> ME" as the subject.
>
> >> --
> >> ~~~~~
> >> Register Today for Cloud Slam 2010 at official website -
> >>http://cloudslam10.com
> >> Posting guidelines:

> >>http://groups.google.ca/group/cloud-computing/web/frequently-asked-qu...
> >> Follow us on Twitterhttp://twitter.com/cloudcomp_groupor
> >> @cloudcomp_group
> >> Post Job/Resume athttp://cloudjobs.net


> >> Buy 88 conference sessions and panels on cloud computing on DVD at
> >>http://www.amazon.com/gp/product/B002H07SEC,

> >>http://www.amazon.com/gp/product/B002H0IW1Uor get instant access to


> >> downloadable versions at
> >>http://cloudslam09.com/content/registration-5.html
>
> >> ~~~~~
> >> You received this message because you are subscribed to the Google Groups
> >> "Cloud Computing" group.
> >> To post to this group, send email to cloud-c...@googlegroups.com
> >> To unsubscribe from this group, send email to
> >> cloud-computi...@googlegroups.com
>
> >> To unsubscribe from this group, send email to cloud-computing+
> >> unsubscribegooglegroups.com or reply to this email with the words "REMOVE
> >> ME" as the subject.
>
> >> --
> >> ~~~~~
> >> Register Today for Cloud Slam 2010 at official website -
> >>http://cloudslam10.com
> >> Posting guidelines:

> >>http://groups.google.ca/group/cloud-computing/web/frequently-asked-qu...
> >> Follow us on Twitterhttp://twitter.com/cloudcomp_groupor
> >> @cloudcomp_group
> >> Post Job/Resume athttp://cloudjobs.net


> >> Buy 88 conference sessions and panels on cloud computing on DVD at
> >>http://www.amazon.com/gp/product/B002H07SEC,
>

> ...
>
> read more »

Ray Nugent

unread,
Apr 2, 2010, 12:53:03 AM4/2/10
to cloud-c...@googlegroups.com
Thanks for the great question Greg. Our patents are in the pipeline right now so I have to be non-specific about the scheme but it's really not rocket science. At a high level one way to achieve the same end result is to describe the workload in some kind of meta data or DSL, reposition that definition in the target cloud or location and xfer the database in the background ahead of the scheduled move. Then switch. Yes, you'll miss a few transactions but that can be dealt with too.

If you think about the notion of moving a live, running VM you have to think about how many bits you can realistically move at any one time. If the use case is for scheduled maintenance of a single server that is NOT a DB server or is running a small server then moving the bits is do'able and in fact vMotion can do this today up to 200km. But if you expand the use case to move more than that, it get's ugly quickly. What if you need to move a terabyte DB server? Or need to move it to a cloud with a different VM format. Or move over the WAN? So live motion becomes useful in a small number of "move" type scenarios.

Sorry for not being more detailed but I think you get the jist of the thinking :-)

Ray




From: Greg Pfister <greg.p...@gmail.com>
To: Cloud Computing <cloud-c...@googlegroups.com>
Sent: Thu, April 1, 2010 6:12:34 PM
Subject: [ Cloud Computing ] Re: sharing an article on InterCloud - Application scaling across multiple Clouds
Post Job/Resume at http://cloudjobs.net

Buy 88 conference sessions and panels on cloud computing on DVD at


~~~~~
You received this message because you are subscribed to the Google Groups "Cloud Computing" group.
To post to this group, send email to cloud-c...@googlegroups.com
To unsubscribe from this group, send email to cloud-computi...@googlegroups.com

To unsubscribe, reply using "remove me" as the subject.

Greg Pfister

unread,
Apr 2, 2010, 11:08:54 PM4/2/10
to Cloud Computing
Thanks. I think I get it -- sounds like you're leveraging having some
sort of dispatcher in front that can switch a stream of transactions
from box A to box B at the appointed time. And some funniness on the
back end with storage (database).

That will work with the appropriate kind of workload, and yes, it
might be simpler than *motion, depending on the circumstances.

Greg Pfister
http://perilsofparallel.blogspot.com/

On Apr 1, 10:53 pm, Ray Nugent <rnug...@yahoo.com> wrote:
> Thanks for the great question Greg. Our patents are in the pipeline right now so I have to be non-specific about the scheme but it's really not rocket science. At a high level one way to achieve the same end result is to describe the workload in
> some kind of meta data or DSL, reposition that definition in the target cloud or location and xfer the
> database in the background ahead of the scheduled move. Then switch. Yes, you'll miss a few transactions but that can be dealt with too.
>
> If you think about the notion of moving a live, running VM you have to think about how many bits you can realistically move at any one time. If the use case is for scheduled maintenance of a single server that is NOT a DB server or is running a small server then moving the bits is do'able and in fact vMotion can do this today up to 200km. But if you expand the use case to move more than that, it get's ugly quickly. What if you need to move a terabyte DB server? Or need to move it to a cloud with a different VM format. Or move over the WAN? So live motion becomes useful in a small number of "move" type scenarios.
>
> Sorry for not being more detailed but I think you get the jist of the thinking :-)
>
> Ray
>
> ________________________________

> From: Greg Pfister <greg.pfis...@gmail.com>


> To: Cloud Computing <cloud-c...@googlegroups.com>
> Sent: Thu, April 1, 2010 6:12:34 PM
> Subject: [ Cloud Computing ] Re: sharing an article on InterCloud -  Application scaling across multiple Clouds
>
> Not to dispute your point, but could you say how you achieve the same
> end goal or goals without live migration? For example, avoiding
> downtime due to scheduled hardware maintenance?
>

> Greg Pfisterhttp://perilsofparallel.blogspot.com/

> > >>http://www.amazon.com/gp/product/B002H0IW1Uorget instant access to


> > >> downloadable versions at
> > >>http://cloudslam09.com/content/registration-5.html
>
> > >> ~~~~~
> > >> You received this message because you
>

> ...
>
> read more »

Reply all
Reply to author
Forward
0 new messages