Yes, that people keep saying "more expensive on-site approach", which is rarely the case.
EC2, for example, is substantially more expensive than any on-site approach, and still requires all the same IT staff since EC2 offers almost no offloading of IT resources. By almost every measure it's significantly more expensive (except in peripheral education and research cases).
For hosting providers and mini-clouds, these can usually be less expensive than the typical managed services offering, but still more expensive than an on-site solution. It all depends on your on-site needs and scale.
In any case, cost does not appear to be a determining factor, at least today.
And then SLA is a joke - EC2 has an SLA that would make even the most accepting legal department run away with their hand in the air.
From: cloud-computing@googlegroups.com [mailto:cloud-computing@googlegroups.com] On Behalf Of R Chishty
Sent: Monday, January 26, 2009 12:47 AM
To: cloud-computing@googlegroups.com
Subject: [ Cloud Computing ] Re: Is "Internal Cloud" an oxymoron ?
This brings up a interesting area from both Ricky and Miha discussion
One of the challenges that I am seeing in Asia is that people are finding it hard to understand the true nature of cloud computing and the impact both from a user and a business perspective.
This is especially around costing and SLA - sometimes, I believe it is not the technology but the leaders of business that will fail to endorse cloud computing as an alternative to the more expensive on-site approach.
> To: cloud-computing@googlegroups.com
> Subject: [ Cloud Computing ] Re: Is "Internal Cloud" an oxymoron ?
> From: mij...@sbcglobal.net
> Date: Mon, 26 Jan 2009 04:56:34 +0000
> Ricky, to have a real cloud, the users concerns you describe below should be the provider's concern.
> A user shouldn't have any concerns other than how much it costs, assuming he uses the right cloud provider.
> Miha
> Sent from my Verizon Wireless BlackBerry
> -----Original Message-----
> From: Ricky Ho <rickyphyl...@yahoo.com>
> Date: Sun, 25 Jan 2009 16:46:42
> To: <cloud-computing@googlegroups.com>
> Subject: [ Cloud Computing ] Re: Is "Internal Cloud" an oxymoron ?
> IMO, the "definition" of a buzzword is not very interesting and usually it is overloaded and can mean different things to different people. Look at what "characteristics" that the cloud brings may be more interesting. And there are 2 different points of view: the user and the provider.
> User's concern:
> ============
> - Total cost of ownership
> - Speed to deploy/undeploy infrastructure resources to meet changing business workload
> - Data sensitivity protection and audit compliance
> - Reliability and QoS guarantee
> Provider's concern:
> ===============
> - Efficient use of underlying infrastructure, usually achieved by physical sharing with logical isolation between tenants ("Virtualization" is a common means to achieve that).
> - Efficient resource scheduling tool that can handle dynamic regrouping of resources based on temporary usage patterns.
> - Efficient management tools that can keep track of the health, ease the debugging in a highly distributed and virtualized environment.
> Note that I haven't said whether the provider is the in-house IT or a public cloud provider because I don't think this is an important distinction. I also believe "enterprise cloud" will be competing with "public cloud" for a while. And a public cloud will always have the advantage of economy of scale and operation expertise while an enterprise cloud will always have the advantage when data sensitivity is a concern.
> I also think reality of cloud computing for large enterprise is always a hybrid environment. It will be very useful to get some guidelines from this expert group on what should be put on the public cloud versus enterprise cloud, and how does components living in different domains can communicate with each other in a secure way.
> Rgds,
> Ricky
> ----- Original Message ----
> From: Frank D. Greco <fgr...@javasig.com>
> To: cloud-computing@googlegroups.com
> Sent: Sunday, January 25, 2009 2:10:47 PM
> Subject: [ Cloud Computing ] Re: Is "Internal Cloud" an oxymoron ?
> At 05:10 PM 1/24/2009, natisha...@gmail.com wrote:
>>In my view what will see in the near future is hybrid-cloud i.e. local-
>>IT+ Internet Cloud. A typical scenario would be to run testing on
>>public cloud and production internally. There are many other scenarios
>>that jutifies this type of hibrid model.
> Totally agree Nati. In the long-term *perhaps* production systems for
> large enterprises will move to an external cloud. In the near/middle
> term, test/dev environments are ripe for external clouds as long as
> the proper precautions about data security are addressed.
> At a large investment bank (remember those?) several years ago,
> I was responsible for moving components of a fixed income analytics
> system to an external data center run by a large well-known
> systems company. This was done to address over-capacity usage.
> Effectively it was a primitive "inside-the-firewall cloud", ie,
> low-ceiling clouds... ;)
> Much of the focus was on: reducing the network/data latency
> as much as possible, software licenses for the cloud machines
> and last, but definitely not least, was data security.
> My former employer was paranoid about data security for good reason.
> I'm sure other large corporations would have the same stance.
> Part of our agreements with our external customers mandated the
> data would never be "co-mingled" (actual lawyer verbiage in our
> contract) with other customers on the same physical machine. This
> defeats much of the economies of scale advantages of the
> IaaS provider.
> Cloud providers should probably avoid this type of agreement
> in their contracts.
> In the end, we had our external site working fine. However, the costs
> were *very* high. Network plumbing, data security products,
> renegotiating software licenses, telecom provider costs, replication
> of data to the external site (to minimize latency) and not
> to mention the additional cost of the infrastructure-service provider.
> Just some anecdotal info to add to the discussion...
>>The real question about private cloud is whether local-IT would ever
>>be able to compete with the cost and level of service of those of the
>>public cloud.
> Many corporations believe that their spin on IT is what differentiates
> them from their competition. One can argue that this philosophy
> was the driving force behind the creation of cloud services in the
> first place. Would Google and Amazon have used an IaaS service?
> It was their belief that they could do it better than anyone.
> I first heard about utility (ie, cloud) computing from Bill Joy
> in the late 90's. Even Bill said some large companies will continue
> to use their IT as a competitive weapon. But he did say a large
> segment of businesses would leverage utility computing. Bill Joy
> is rarely wrong... ok... the invention of the c-shell wasn't
> exactly that great... ;+
>>This is where the economy of scales plays an important role (See James
>>Liddle recent blog: The Economics of Cloud
>>http://vehera.jsn-server7.com/LiddleBlog/?p=234 >>) i.e. the bigger the data-center is the higher the chance the cost
>>and level of service will get better then those of smaller data-
>>centers.
> Having a very large data center means you need another large
> data center
> as a DR site. It may be better to have families of data
> centers working
> in a federation. It will be interesting to see how well Sun
> Microsystems
> manages their Q-Layer acquisition and virtual data-centers notion.
> We can draw analogies to the power-grid, but at a certain point, this
> analogy doesn't apply. Sort of like the common "a GUI is just like
> a car dashboard" and "building software is like building a house"...
> yeah it is... but yeah it isn't...
>> It may very well that over time public cloud will become the
>>only option just like energy, however its going to take a while till
>>this will happen if at all.
> Totally agreed Nati. I'm not 100% convinced it will be the
> only option
> for all companies in the long run.
> The difference between cloud-based IaaS and the energy
> grid is software. Software innovation has a *long* way to
> go before its
> mature like our energy infrastructure. Besides, large
> public cloud utilities
> eventually will act like internal IT when many customers
> share the same
> virtual/physical infrastructure. :)
For large enterprises, the compelling usage of CC in my opinion is "occasional usage". Such as ... buffer for workload spike, experimenting new ideas, or disaster recovery. They shouldn't expect cost saving by outsourcing their stable workload handling to a cloud provider.
For startups, CC is very compelling so they don't need to pay the infrastructure setup cost at the beginning.
Rgds,
Ricky
________________________________
From: "Pietrasanta, Mark" <Mark.Pietrasa...@aquilent.com>
To: "cloud-computing@googlegroups.com" <cloud-computing@googlegroups.com>
Sent: Monday, January 26, 2009 6:30:34 AM
Subject: [ Cloud Computing ] Re: Is "Internal Cloud" an oxymoron ?
Yes, that people keep saying “more expensive on-site approach”,
which is rarely the case.
EC2, for example, is substantially more expensive than any
on-site approach, and still requires all the same IT staff since EC2 offers
almost no offloading of IT resources. By almost every measure it’s significantly
more expensive (except in peripheral education and research cases).
For hosting providers and mini-clouds, these can usually be less
expensive than the typical managed services offering, but still more expensive
than an on-site solution. It all depends on your on-site needs and scale.
In any case, cost does not appear to be a determining factor, at
least today.
And then SLA is a joke – EC2 has an SLA that would make even the
most accepting legal department run away with their hand in the air.
From:cloud-computing@googlegroups.com
[mailto:cloud-computing@googlegroups.com] On Behalf Of R Chishty
Sent: Monday, January 26, 2009 12:47 AM
To: cloud-computing@googlegroups.com
Subject: [ Cloud Computing ] Re: Is "Internal Cloud" an
oxymoron ?
This
brings up a interesting area from both Ricky and Miha discussion
One
of the challenges that I am seeing in Asia is that people are finding it hard
to understand the true nature of cloud computing and the impact both from a
user and a business perspective.
This
is especially around costing and SLA - sometimes, I believe it is not the
technology but the leaders of business that will fail to endorse cloud
computing as an alternative to the more expensive on-site approach.
Any
comments?
Kind
Regards
Ruz
Triactive.asia
> To: cloud-computing@googlegroups.com
> Subject: [ Cloud Computing ] Re: Is "Internal Cloud" an oxymoron
?
> From: mij...@sbcglobal.net
> Date: Mon, 26 Jan 2009 04:56:34 +0000
> Ricky, to have a real cloud, the users concerns you describe below should
be the provider's concern.
> A user shouldn't have any concerns other than how much it costs, assuming
> -----Original Message-----
> From: Ricky Ho <rickyphyl...@yahoo.com>
> Date: Sun, 25 Jan 2009 16:46:42 > To: <cloud-computing@googlegroups.com>
> Subject: [ Cloud Computing ] Re: Is "Internal Cloud" an oxymoron
?
> IMO, the "definition" of a buzzword is not very interesting and
usually it is overloaded and can mean different things to different people.
Look at what "characteristics" that the cloud brings may be more
interesting. And there are 2 different points of view: the user and the
provider.
> User's concern:
> ============
> - Total cost of ownership
> - Speed to deploy/undeploy infrastructure resources to meet changing
business workload
> - Data sensitivity protection and audit compliance
> - Reliability and QoS guarantee
> Provider's concern:
> ===============
> - Efficient use of underlying infrastructure, usually achieved by physical
sharing with logical isolation between tenants ("Virtualization" is a
common means to achieve that).
> - Efficient resource scheduling tool that can handle dynamic regrouping of
resources based on temporary usage patterns.
> - Efficient management tools that can keep track of the health, ease the
debugging in a highly distributed and virtualized environment.
> Note that I haven't said whether the provider is the in-house IT or a
public cloud provider because I don't think this is an important distinction. I
also believe "enterprise cloud" will be competing with "public
cloud" for a while. And a public cloud will always have the advantage of
economy of scale and operation expertise while an enterprise cloud will always
have the advantage when data sensitivity is a concern.
> I also think reality of cloud computing for large enterprise is always a
hybrid environment. It will be very useful to get some guidelines from this
expert group on what should be put on the public cloud versus enterprise cloud,
and how does components living in different domains can communicate with each
other in a secure way.
> ----- Original Message ----
> From: Frank D. Greco <fgr...@javasig.com>
> To: cloud-computing@googlegroups.com
> Sent: Sunday, January 25, 2009 2:10:47 PM
> Subject: [ Cloud Computing ] Re: Is "Internal Cloud" an oxymoron
?
> At 05:10 PM 1/24/2009, natisha...@gmail.com wrote:
>>In my view what will see in the near future is hybrid-cloud i.e. local-
>>IT+ Internet Cloud. A typical scenario would be to run testing on
>>public cloud and production internally. There are many other scenarios
>>that jutifies this type of hibrid model.
> Totally agree Nati. In the long-term *perhaps* production systems for
> large enterprises will move to an external cloud. In the near/middle
> term, test/dev environments are ripe for external clouds as long as
> the proper precautions about data security are addressed.
> At a large investment bank (remember those?) several years ago,
> I was responsible for moving components of a fixed income analytics
> system to an external data center run by a large well-known
> systems company. This was done to address over-capacity usage.
> Effectively it was a primitive "inside-the-firewall cloud", ie,
> low-ceiling clouds... ;)
> Much of the focus was on: reducing the network/data latency
> as much as possible, software licenses for the cloud machines
> and last, but definitely not least, was data security.
> My former employer was paranoid about data security for good reason.
> I'm sure other large corporations would have the same stance.
> Part of our agreements with our external customers mandated the
> data would never be "co-mingled" (actual lawyer verbiage in our
> contract) with other customers on the same physical machine. This
> defeats much of the economies of scale advantages of the > IaaS provider.
> Cloud providers should probably avoid this type of agreement
> in their contracts.
> In the end, we had our external site working fine. However, the costs
> were *very* high. Network plumbing, data security products,
> renegotiating software licenses, telecom provider costs, replication
> of data to the external site (to minimize latency) and not
> to mention the additional cost of the infrastructure-service provider.
> Just some anecdotal info to add to the discussion...
>>The real question about private cloud is whether local-IT would ever
>>be able to compete with the cost and level of service of those of the
>>public cloud.
> Many corporations believe that their spin on IT is what differentiates
> them from their competition. One can argue that this philosophy
> was the driving force behind the creation of cloud services in the
> first place. Would Google and Amazon have used an IaaS service?
> It was their belief that they could do it better than anyone.
> I first heard about utility (ie, cloud) computing from Bill Joy
> in the late 90's. Even Bill said some large companies will continue
> to use their IT as a competitive weapon. But he did say a large
> segment of businesses would leverage utility computing. Bill Joy
> is rarely wrong... ok... the invention of the c-shell wasn't
> exactly that great... ;+
>>This is where the economy of scales plays an important role (See James
>>Liddle recent blog: The Economics of Cloud >>http://vehera.jsn-server7.com/LiddleBlog/?p=234 >>) i.e. the bigger the data-center is the higher the chance the cost
>>and level of service will get better then those of smaller data-
>>centers.
> Having a very large data center means you need another large > data center
> as a DR site. It may be better to have families of data > centers working
> in a federation. It will be interesting to see how well Sun > Microsystems
> manages their Q-Layer acquisition and virtual data-centers notion.
> We can draw analogies to the power-grid, but at a certain point, this
> analogy doesn't apply. Sort of like the common "a GUI is just like
> a car dashboard" and "building software is like building a
house"...
> yeah it is... but yeah it isn't...
>> It may very well that over time public cloud will become the
>>only option just like energy, however its going to take a while till
>>this will happen if at all.
> Totally agreed Nati. I'm not 100% convinced it will be the > only option
> for all companies in the long run.
> The difference between cloud-based IaaS and the energy
> grid is software. Software innovation has a *long* way to > go before its
> mature like our energy infrastructure. Besides, large > public cloud utilities
> eventually will act like internal IT when many customers > share the same
> virtual/physical infrastructure. :)
Ricky - let me build on your idea of experimentation.
Business managers like to experiment. Ie lets try offering bonzoberry
flavored baby formula in the Cleveland market. If it sells, we'll launch
regionally next this fall, then nationally next spring. This is call a "low
cost probe" and it typically rolls out fast. But, these types of things
have a high failure rate. They want to try it, and kill it if not working.
What I see is that IT and business typically have different time scales. IT
likes more certainty in projected capacity, connectivity etc in order to
scope an effective solution, build, test, provision to a realistic set of
requirements.
In a recent studied I conducted, I found that top performing enterprise IT
shops seem to be better at killing projects that no longer make business
sense. I assume these folks are better aligned with business on low cost
probes.
In another recent study, I found that enterprise IT often has project office
that manages more strategic big projects, but often struggles with managing
ongoing stream of more tactical requests. Ie - we need 3 more servers, we
need more storage etc. I found evidence that having a defined and published
process for handling the tactical demand is actually a sign of the highest
level of governance maturity.
My gut is that CC offers potential value for internal IT to help facilitate
the smaller, faster low cost probe type projects via plug in resources.
Perhaps once the business need is proven and scoped based on measured usage,
then IT can find a more permanent home (internal or external) based on more
stable set of requirements/projections.
ON a side note -
Per other recent posts and some conversations I had last week with
infrastructure server providers, I do think that there is a layer of
services IT organizations need to offer to manage CC based projects. My
guess is that business funders/managers will still turn to IT to help manage
CC based capabilities.
Ricky Ho posted a comment about list of functions the abstraction layer
provides.
We should also consider what "services" the IT staff can offer, as well as
technical services the CC management infrastructure should offer.
Kurt
_____
From: cloud-computing@googlegroups.com
[mailto:cloud-computing@googlegroups.com] On Behalf Of Ricky Ho
Sent: Monday, January 26, 2009 8:35 AM
To: cloud-computing@googlegroups.com
Subject: [ Cloud Computing ] Re: Is "Internal Cloud" an oxymoron ?
I agree.
For large enterprises, the compelling usage of CC in my opinion is
"occasional usage". Such as ... buffer for workload spike, experimenting
new ideas, or disaster recovery. They shouldn't expect cost saving by
outsourcing their stable workload handling to a cloud provider.
For startups, CC is very compelling so they don't need to pay the
infrastructure setup cost at the beginning.
Rgds,
Ricky
_____
From: "Pietrasanta, Mark" <Mark.Pietrasa...@aquilent.com>
To: "cloud-computing@googlegroups.com" <cloud-computing@googlegroups.com>
Sent: Monday, January 26, 2009 6:30:34 AM
Subject: [ Cloud Computing ] Re: Is "Internal Cloud" an oxymoron ?
Yes, that people keep saying "more expensive on-site approach", which is
rarely the case.
EC2, for example, is substantially more expensive than any on-site approach,
and still requires all the same IT staff since EC2 offers almost no
offloading of IT resources. By almost every measure it's significantly more
expensive (except in peripheral education and research cases).
For hosting providers and mini-clouds, these can usually be less expensive
than the typical managed services offering, but still more expensive than an
on-site solution. It all depends on your on-site needs and scale.
In any case, cost does not appear to be a determining factor, at least
today.
And then SLA is a joke - EC2 has an SLA that would make even the most
accepting legal department run away with their hand in the air.
From: cloud-computing@googlegroups.com
[mailto:cloud-computing@googlegroups.com] On Behalf Of R Chishty
Sent: Monday, January 26, 2009 12:47 AM
To: cloud-computing@googlegroups.com
Subject: [ Cloud Computing ] Re: Is "Internal Cloud" an oxymoron ?
This brings up a interesting area from both Ricky and Miha discussion
One of the challenges that I am seeing in Asia is that people are finding it
hard to understand the true nature of cloud computing and the impact both
from a user and a business perspective.
This is especially around costing and SLA - sometimes, I believe it is not
the technology but the leaders of business that will fail to endorse cloud
computing as an alternative to the more expensive on-site approach.
Any comments?
Kind Regards
Ruz
Triactive.asia
> To: cloud-computing@googlegroups.com
> Subject: [ Cloud Computing ] Re: Is "Internal Cloud" an oxymoron ?
> From: mij...@sbcglobal.net
> Date: Mon, 26 Jan 2009 04:56:34 +0000
> Ricky, to have a real cloud, the users concerns you describe below should
be the provider's concern.
> A user shouldn't have any concerns other than how much it costs, assuming
he uses the right cloud provider.
> Miha
> Sent from my Verizon Wireless BlackBerry
> -----Original Message-----
> From: Ricky Ho <rickyphyl...@yahoo.com>
> Date: Sun, 25 Jan 2009 16:46:42 > To: <cloud-computing@googlegroups.com>
> Subject: [ Cloud Computing ] Re: Is "Internal Cloud" an oxymoron ?
> IMO, the "definition" of a buzzword is not very interesting and usually it
is overloaded and can mean different things to different people. Look at
what "characteristics" that the cloud brings may be more interesting. And
there are 2 different points of view: the user and the provider.
> User's concern:
> ============
> - Total cost of ownership
> - Speed to deploy/undeploy infrastructure resources to meet changing
business workload
> - Data sensitivity protection and audit compliance
> - Reliability and QoS guarantee
> Provider's concern:
> ===============
> - Efficient use of underlying infrastructure, usually achieved by physical
sharing with logical isolation between tenants ("Virtualization" is a common
means to achieve that).
> - Efficient resource scheduling tool that can handle dynamic regrouping of
resources based on temporary usage patterns.
> - Efficient management tools that can keep track of the health, ease the
debugging in a highly distributed and virtualized environment.
> Note that I haven't said whether the provider is the in-house IT or a
public cloud provider because I don't think this is an important
distinction. I also believe "enterprise cloud" will be competing with
"public cloud" for a while. And a public cloud will always have the
advantage of economy of scale and operation expertise while an enterprise
cloud will always have the advantage when data sensitivity is a concern.
> I also think reality of cloud computing for large enterprise is always a
hybrid environment. It will be very useful to get some guidelines from this
expert group on what should be put on the public cloud versus enterprise
cloud, and how does components living in different domains can communicate
with each other in a secure way.
> ----- Original Message ----
> From: Frank D. Greco <fgr...@javasig.com>
> To: cloud-computing@googlegroups.com
> Sent: Sunday, January 25, 2009 2:10:47 PM
> Subject: [ Cloud Computing ] Re: Is "Internal Cloud" an oxymoron ?
> At 05:10 PM 1/24/2009, natisha...@gmail.com wrote:
>>In my view what will see in the near future is hybrid-cloud i.e. local-
>>IT+ Internet Cloud. A typical scenario would be to run testing on
>>public cloud and production internally. There are many other scenarios
>>that jutifies this type of hibrid model.
> Totally agree Nati. In the long-term *perhaps* production systems for
> large enterprises will move to an external cloud. In the near/middle
> term, test/dev environments are ripe for external clouds as long as
> the proper precautions about data security are addressed.
> At a large investment bank (remember those?) several years ago,
> I was responsible for moving components of a fixed income analytics
> system to an external data center run by a large well-known
> systems company. This was done to address over-capacity usage.
> Effectively it was a primitive "inside-the-firewall cloud", ie,
> low-ceiling clouds... ;)
> Much of the focus was on: reducing the network/data latency
> as much as possible, software licenses for the cloud machines
> and last, but definitely not least, was data security.
> My former employer was paranoid about data security for good reason.
> I'm sure other large corporations would have the same stance.
> Part of our agreements with our external customers mandated the
> data would never be "co-mingled" (actual lawyer verbiage in our
> contract) with other customers on the same physical machine. This
> defeats much of the economies of scale advantages of the > IaaS provider.
> Cloud providers should probably avoid this type of agreement
> in their contracts.
> In the end, we had our external site working fine. However, the costs
> were *very* high. Network plumbing, data security products,
> renegotiating software licenses, telecom provider costs, replication
> of data to the external site (to minimize latency) and not
> to mention the additional cost of the infrastructure-service provider.
> Just some anecdotal info to add to the discussion...
>>The real question about private cloud is whether local-IT would ever
>>be able to compete with the cost and level of service of those of the
Discussion subject changed to "The cloud ideas from this group. Can they be funded by VCs? RE: [ Cloud Computing ] Re: Is "Internal Cloud" an oxymoron ?" by Miha Ahronovitz
I touch the subject of funding inspired by the this post from Rick.
Rick, you state as user concerns the provider should consider
- Speed to deploy/undeploy infrastructure resources to meet changing
business workload
- Data sensitivity protection and audit compliance
- Reliability and QoS guarantee
And earlier you mentioned:
- Total cost of ownership
As Product Manager, I have to translate these requirements in a language the
VCs (Venture Capitalists) understand.
I learned "the little bird" methodology from Guy Kawasaki book: "The Art of the Start: The Time-Tested, Battle-Hardened Guide for Anyone
Starting Anything"
Each time I make a statement, a little bird sits on my shoulder and asks:
"So What?" See my answers to the little bird.
1 Bullet Total cost of ownership
little bird: So What?
Translation: This is complicated to calculate and not obvious. Clouds make
profits, they are not cost centers. We need special research to find a
number, and then, skeptics will not believe. The TCO is of interest, it
adopted, to providers only. Users only want to know a price list for
services. They shop around and do their own calculations that the provider
cannot make for them. They will not order an internal TCO before making a
decision. They will look at their budgets and see how to get the maximum for
each buck
2. Bullet
- Speed to deploy/undeploy infrastructure resources to meet changing
business workload
little bird: So What?
Translation: This assumes the user has a say in deploying resources. If he does, then
s/he does not use the cloud, but simply an external provider. If it is a
cloud, I can see the provider has always an answer. The user just asks how
long it takes to accommodate x simultaneous users and y applications
delivered as services, at QOS defined formally by Service Level Agreement.
The resources , ideally should assigned dynamically to the customer, in such
a way s/he is totally unaware of. All the resources consumed are part of
the price quoted next to the SLA.
3. Bullet:
- Data sensitivity protection and audit compliance
little bird: So What?
Translation:
There is no way to assure for everyone that their data is "safe" in a
particular cloud. People have different degrees of data protection. If there
is special request for high level audit needs and data protection, this
should be priced separately. If it cannot be offered, then the user better
run these services internally. The cloud owner will say: This is what I can
offer. Do you accept the agreement or not?
4. Bullet
- Reliability and QoS guarantee
little bird: So What?
Translation:
Customers and cloud providers agree on quality of service that should be
priced. They write a Service Level Agreement (SLA) and this should be clear
enough for each requirement. Each line in the SLA should be associated (by
the Cloud Provider) to a Service Level Objectives (SLOs). Which are just
numbers. The provider understands, the user understands, they sign and all
what is needed. Even simpler, the providers offers prices for various levels
of service assurance, and the user accepts the range that it finds
appropriate.
There a few slides missing. If I start a cloud as above, who will be my
customers (they want NAMES , not just a market definition). Why these
customers will select my services versus anything else (product , service)
that solved for the customer the same problem? (Because I faster, cheaper,
more reliable, easier to adopt than the others?) etc.
Last paragraph raises the question: Do we worry on this group whether our
cloud ideas can be funded? Do we worry on this group whether all this
technology ca marketed and transformed in $?
This is why entrepreneurial Product Management exists. To make a business in
the eyes of the investors and to mitigate all risks (translation; to answer
very hard questions) to an acceptable level to the investors.
Cheers,
Miha
PS: Now I turn away my voice and hope that the bird does not hear me :-) To
learn more about how SLAs and SLOs work in Sun Grid Engine with the Service
Domain Manager, here it is a detailed engineering presentation by my
colleague Richard Hierlmeier:
[mailto:cloud-computing@googlegroups.com] On Behalf Of Ricky Ho
Sent: Sunday, January 25, 2009 10:02 PM
To: cloud-computing@googlegroups.com
Subject: [ Cloud Computing ] Re: Is "Internal Cloud" an oxymoron ?
Sorry that I didn't mention it explicitly. I thought it is obvious that the
provider's concern is to address the user's concern, in order to stay
competitive.
I also think other than cost, the users are concerning about ...
- Speed to deploy/undeploy infrastructure resources to meet changing
business workload
- Data sensitivity protection and audit compliance
- Reliability and QoS guarantee
A provider which is cheap but not providing the above is not very
attractive. Right ?
Rgds,
Ricky
----- Original Message ----
From: "mij...@sbcglobal.net" <mij...@sbcglobal.net>
To: cloud-computing@googlegroups.com
Sent: Sunday, January 25, 2009 8:56:34 PM
Subject: [ Cloud Computing ] Re: Is "Internal Cloud" an oxymoron ?
Ricky, to have a real cloud, the users concerns you describe below should be
the provider's concern.
A user shouldn't have any concerns other than how much it costs, assuming
he uses the right cloud provider.
Miha
Sent from my Verizon Wireless BlackBerry
-----Original Message-----
From: Ricky Ho <rickyphyl...@yahoo.com>
Date: Sun, 25 Jan 2009 16:46:42 To: <cloud-computing@googlegroups.com>
Subject: [ Cloud Computing ] Re: Is "Internal Cloud" an oxymoron ?
IMO, the "definition" of a buzzword is not very interesting and usually it
is overloaded and can mean different things to different people. Look at
what "characteristics" that the cloud brings may be more interesting. And
there are 2 different points of view: the user and the provider.
User's concern:
============
- Total cost of ownership
- Speed to deploy/undeploy infrastructure resources to meet changing
business workload
- Data sensitivity protection and audit compliance
- Reliability and QoS guarantee
Provider's concern:
===============
- Efficient use of underlying infrastructure, usually achieved by physical
sharing with logical isolation between tenants ("Virtualization" is a common
means to achieve that).
- Efficient resource scheduling tool that can handle dynamic regrouping of
resources based on temporary usage patterns.
- Efficient management tools that can keep track of the health, ease the
debugging in a highly distributed and virtualized environment.
Note that I haven't said whether the provider is the in-house IT or a public
cloud provider because I don't think this is an important distinction. I
also believe "enterprise cloud" will be competing with "public cloud" for a
while. And a public cloud will always have the advantage of economy of
scale and operation expertise while an enterprise cloud will always have the
advantage when data sensitivity is a concern.
I also think reality of cloud computing for large enterprise is always a
hybrid environment. It will be very useful to get some guidelines from this
expert group on what should be put on the public cloud versus enterprise
cloud, and how does components living in different domains can communicate
with each other in a secure way.
Rgds,
Ricky
----- Original Message ----
From: Frank D. Greco <fgr...@javasig.com>
To: cloud-computing@googlegroups.com
Sent: Sunday, January 25, 2009 2:10:47 PM
Subject: [ Cloud Computing ] Re: Is "Internal Cloud" an oxymoron ?
At 05:10 PM 1/24/2009, natisha...@gmail.com wrote:
>In my view what will see in the near future is hybrid-cloud i.e. local-
>IT+ Internet Cloud. A typical scenario would be to run testing on
>public cloud and production internally. There are many other scenarios
>that jutifies this type of hibrid model.
Totally agree Nati. In the long-term *perhaps* production systems
for
large enterprises will move to an external cloud. In the
near/middle
term, test/dev environments are ripe for external clouds as long as
the proper precautions about data security are addressed.
At a large investment bank (remember those?) several years ago,
I was responsible for moving components of a fixed income analytics
system to an external data center run by a large well-known
systems company. This was done to address over-capacity usage.
Effectively it was a primitive "inside-the-firewall cloud", ie,
low-ceiling clouds... ;)
Much of the focus was on: reducing the network/data latency
as much as possible, software licenses for the cloud machines
and last, but definitely not least, was data security.
My former employer was paranoid about data security for good
reason.
I'm sure other large corporations would have the same stance.
Part of our agreements with our external customers mandated the
data would never be "co-mingled" (actual lawyer verbiage in our
contract) with other customers on the same physical machine. This
defeats much of the economies of scale advantages of the IaaS provider.
Cloud providers should probably avoid this type of agreement
in their contracts.
In the end, we had our external site working fine. However, the
costs
were *very* high. Network plumbing, data security products,
renegotiating software licenses, telecom provider costs,
From: cloud-computing@googlegroups.com [mailto:cloud-computing@googlegroups.com] On Behalf Of Kurt Milne
Sent: Monday, January 26, 2009 11:42 AM
To: cloud-computing@googlegroups.com
Subject: [ Cloud Computing ] Re: Is "Internal Cloud" an oxymoron ?
Ricky - let me build on your idea of experimentation.
Business managers like to experiment. Ie lets try offering bonzoberry flavored baby formula in the Cleveland market. If it sells, we'll launch regionally next this fall, then nationally next spring. This is call a "low cost probe" and it typically rolls out fast. But, these types of things have a high failure rate. They want to try it, and kill it if not working.
What I see is that IT and business typically have different time scales. IT likes more certainty in projected capacity, connectivity etc in order to scope an effective solution, build, test, provision to a realistic set of requirements.
In a recent studied I conducted, I found that top performing enterprise IT shops seem to be better at killing projects that no longer make business sense. I assume these folks are better aligned with business on low cost probes.
In another recent study, I found that enterprise IT often has project office that manages more strategic big projects, but often struggles with managing ongoing stream of more tactical requests. Ie - we need 3 more servers, we need more storage etc. I found evidence that having a defined and published process for handling the tactical demand is actually a sign of the highest level of governance maturity.
My gut is that CC offers potential value for internal IT to help facilitate the smaller, faster low cost probe type projects via plug in resources. Perhaps once the business need is proven and scoped based on measured usage, then IT can find a more permanent home (internal or external) based on more stable set of requirements/projections.
ON a side note -
Per other recent posts and some conversations I had last week with infrastructure server providers, I do think that there is a layer of services IT organizations need to offer to manage CC based projects. My guess is that business funders/managers will still turn to IT to help manage CC based capabilities.
Ricky Ho posted a comment about list of functions the abstraction layer provides.
We should also consider what "services" the IT staff can offer, as well as technical services the CC management infrastructure should offer.
Kurt
________________________________
From: cloud-computing@googlegroups.com [mailto:cloud-computing@googlegroups.com] On Behalf Of Ricky Ho
Sent: Monday, January 26, 2009 8:35 AM
To: cloud-computing@googlegroups.com
Subject: [ Cloud Computing ] Re: Is "Internal Cloud" an oxymoron ?
I agree.
For large enterprises, the compelling usage of CC in my opinion is "occasional usage". Such as ... buffer for workload spike, experimenting new ideas, or disaster recovery. They shouldn't expect cost saving by outsourcing their stable workload handling to a cloud provider.
For startups, CC is very compelling so they don't need to pay the infrastructure setup cost at the beginning.
Rgds,
Ricky
________________________________
From: "Pietrasanta, Mark" <Mark.Pietrasa...@aquilent.com>
To: "cloud-computing@googlegroups.com" <cloud-computing@googlegroups.com>
Sent: Monday, January 26, 2009 6:30:34 AM
Subject: [ Cloud Computing ] Re: Is "Internal Cloud" an oxymoron ?
Yes, that people keep saying "more expensive on-site approach", which is rarely the case.
EC2, for example, is substantially more expensive than any on-site approach, and still requires all the same IT staff since EC2 offers almost no offloading of IT resources. By almost every measure it's significantly more expensive (except in peripheral education and research cases).
For hosting providers and mini-clouds, these can usually be less expensive than the typical managed services offering, but still more expensive than an on-site solution. It all depends on your on-site needs and scale.
In any case, cost does not appear to be a determining factor, at least today.
And then SLA is a joke - EC2 has an SLA that would make even the most accepting legal department run away with their hand in the air.
From: cloud-computing@googlegroups.com [mailto:cloud-computing@googlegroups.com] On Behalf Of R Chishty
Sent: Monday, January 26, 2009 12:47 AM
To: cloud-computing@googlegroups.com
Subject: [ Cloud Computing ] Re: Is "Internal Cloud" an oxymoron ?
This brings up a interesting area from both Ricky and Miha discussion
One of the challenges that I am seeing in Asia is that people are finding it hard to understand the true nature of cloud computing and the impact both from a user and a business perspective.
This is especially around costing and SLA - sometimes, I believe it is not the technology but the leaders of business that will fail to endorse cloud computing as an alternative to the more expensive on-site approach.
> To: cloud-computing@googlegroups.com
> Subject: [ Cloud Computing ] Re: Is "Internal Cloud" an oxymoron ?
> From: mij...@sbcglobal.net
> Date: Mon, 26 Jan 2009 04:56:34 +0000
> Ricky, to have a real cloud, the users concerns you describe below should be the provider's concern.
> A user shouldn't have any concerns other than how much it costs, assuming he uses the right cloud provider.
> Miha
> Sent from my Verizon Wireless BlackBerry
> -----Original Message-----
> From: Ricky Ho <rickyphyl...@yahoo.com>
> Date: Sun, 25 Jan 2009 16:46:42
> To: <cloud-computing@googlegroups.com>
> Subject: [ Cloud Computing ] Re: Is "Internal Cloud" an oxymoron ?
> IMO, the "definition" of a buzzword is not very interesting and usually it is overloaded and can mean different things to different people. Look at what "characteristics" that the cloud brings may be more interesting. And there are 2 different points of view: the user and the provider.
> User's concern:
> ============
> - Total cost of ownership
> - Speed to deploy/undeploy infrastructure resources to meet changing business workload
> - Data sensitivity protection and audit compliance
> - Reliability and QoS guarantee
> Provider's concern:
> ===============
> - Efficient use of underlying infrastructure, usually achieved by physical sharing with logical isolation between tenants ("Virtualization" is a common means to achieve that).
> - Efficient resource scheduling tool that can handle dynamic regrouping of resources based on temporary usage patterns.
> - Efficient management tools that can keep track of the health, ease the debugging in a highly distributed and virtualized environment.
> Note that I haven't said whether the provider is the in-house IT or a public cloud provider because I don't think this is an important distinction. I also believe "enterprise cloud" will be competing with "public cloud" for a while. And a public cloud will always have the advantage of economy of scale and operation expertise while an enterprise cloud will always have the advantage when data sensitivity is a concern.
> I also think reality of cloud computing for large enterprise is always a hybrid environment. It will be very useful to get some guidelines from this expert group on what should be put on the public cloud versus enterprise cloud, and how does components living in different domains can communicate with each other in a secure way.
> Rgds,
> Ricky
> ----- Original Message ----
> From: Frank D. Greco <fgr...@javasig.com>
> To: cloud-computing@googlegroups.com
> Sent: Sunday, January 25, 2009 2:10:47 PM
> Subject: [ Cloud Computing ] Re: Is "Internal Cloud" an oxymoron ?
> At 05:10 PM 1/24/2009, natisha...@gmail.com wrote:
>>In my view what will see in the near future is hybrid-cloud i.e. local-
>>IT+ Internet Cloud. A typical scenario would be to run testing on
>>public cloud and production internally. There are many other scenarios
>>that jutifies this type of hibrid model.
> Totally agree Nati. In the long-term *perhaps* production systems for
> large enterprises will move to an external cloud. In the near/middle
> term, test/dev environments are ripe for external clouds as long as
> the proper precautions about data security are addressed.
> At a large investment bank (remember those?) several years ago,
> I was responsible for moving components of a fixed income analytics
> system to an external data center run by a large well-known
> systems company. This was done to address over-capacity usage.
> Effectively it was a primitive "inside-the-firewall cloud", ie,
> low-ceiling clouds... ;)
> Much of the focus was on: reducing the network/data latency
> as much as possible, software licenses for the cloud machines
> and last, but definitely not least, was data security.
> My former employer was paranoid about data security for good reason.
> I'm sure other large corporations would have the same stance.
> Part of our agreements with our external customers mandated the
> data would never be "co-mingled" (actual lawyer verbiage in our
> contract) with other customers on the same physical machine. This
> defeats much of the economies of scale advantages of the
> IaaS provider.
> Cloud providers should probably avoid this type of agreement
> in their contracts.
> In the end, we had our external site working fine. However, the costs
> were *very* high. Network plumbing, data security products,
> renegotiating software licenses, telecom provider costs, replication
> of data to the external site (to minimize latency) and not
> to mention the additional cost of the infrastructure-service provider.
> Just some anecdotal info to add to the discussion...
>>The real question about private cloud is whether local-IT
For me, Cloud Computing, as a service, is a reincarnation of the
ancient time sharing with updated [distributed systems] technologies.
With time sharing, an organization can set up a time sharing system
for its users (as in the case of practically all the universities) or
contract out the computing need of its users to a time sharing system
company.
As such, I do not find "internal cloud" to be an oxymoron.
Imitation is the sincerest form of flattery. Some guys at the US DoD
see Amazon and other clouds and for some reason they cannot use them
yet. Perhaps naturally they say "I want one of those". As Jim and
Stu pointed out earlier on this thread, most of the economics of
centrally managed, self-serviced, audited multitenancy, carry across
to the 'internal' case. So "mimicry" is the way forward for some
customers.
The corollary of the statement "there can be no internal cloud" is
that EC2 would cease to be a cloud if Amazon made it so that only
Amazon staff and suppliers could use it, but Amazon customers could
not. This does not make sense, therefore the notion of an internal
cloud is viable.
On Tue, Jan 27, 2009 at 7:17 PM, <johnle...@ieee.org> wrote:
> For me, Cloud Computing, as a service, is a reincarnation of the
> ancient time sharing with updated [distributed systems] technologies.
> With time sharing, an organization can set up a time sharing system
> for its users (as in the case of practically all the universities) or
> contract out the computing need of its users to a time sharing system
> company.
> As such, I do not find "internal cloud" to be an oxymoron.
In fact it does make sense that AWS ceases to be a cloud once it is used by a finite audience. At that point they will not be able to jusify growth. They will hit the end of their cloud resource and need to order more. This means adding a machine has gone from an API call to a PO.
Ray
On Jan 27, 2009, at 12:12 PM, Alexis Richardson <alexis.richard...@gmail.com> wrote:
Here is what strikes me as a perfectly valid way of looking at this:
Imitation is the sincerest form of flattery. Some guys at the US DoD
see Amazon and other clouds and for some reason they cannot use them
yet. Perhaps naturally they say "I want one of those". As Jim and
Stu pointed out earlier on this thread, most of the economics of
centrally managed, self-serviced, audited multitenancy, carry across
to the 'internal' case. So "mimicry" is the way forward for some
customers.
The corollary of the statement "there can be no internal cloud" is
that EC2 would cease to be a cloud if Amazon made it so that only
Amazon staff and suppliers could use it, but Amazon customers could
not. This does not make sense, therefore the notion of an internal
cloud is viable.
alexis
On Tue, Jan 27, 2009 at 7:17 PM, <johnle...@ieee.org> wrote:
For me, Cloud Computing, as a service, is a reincarnation of the
ancient time sharing with updated [distributed systems] technologies.
With time sharing, an organization can set up a time sharing system
for its users (as in the case of practically all the universities) or
contract out the computing need of its users to a time sharing system
company.
As such, I do not find "internal cloud" to be an oxymoron.
I have been following some of the chatter in the "Cloud Standards" space. Looks to me, we have more talk about standardization work than actual cloud work. May be I am moving in the wrong circles ;o)
Anyway the reason for the blog is Bob Sutor's notes from a Cloud Standards gathering. To be fair, I did not attend the meeting and I should be careful to separate the message from the messenger - Bob's ideas vs. what he is reporting from the gathering.
A few things that caught my eye - some I agree, a lot I disagree:
== There could be potentially 100s of standards
To quote Scooby-doo, "Yikes". I think this is the wrong approach. The domain (Cloud Computing) is barely born and we are already talking about 100s of standards- new or old ! This will definitely slow down the progress or (most probably) everyone will ignore the "standards" and do whatever is necessary to run the business. We should strive for simplification. I propose that overstandardization was one of the reasons for the demise of SOA and Web Services. (Yep, I know lots of people will disagree. But before you start throwing stones, at least read thru Ann's blog - http://apsblog.burtongroup.com/2009/01/soa-is-dead-long-live-services.ht ml)
== Need "marketecture" & speed of standardization is important
This is exactly where I think we have the cart before the horse. Usually standardization happens based on a few running instances and when practitioners realize that some parts need interoperability based on experience (let me say again, actual experience) standardization happens. Because of running code, protocols et al, the participants know what to standard and how. How are we going to standardize based on an imaginary marketecture ?
And we should never rush standards - usually standards a little deliberate and systemic. Agreed, once we understand a domain, we shouldn't take infinite time. But I have never heard of standardization based on "marketecture" even before a domain is being developed.
== Need to understand interchangeable parts vs. interoperability
I think we are over analyzing here. To tell the truth, I have no clue what it means.
== "It is not acceptable to delay standardization until a particular provider establishes lock-in or a monopoly"
Eh ? ;o) Very strange. Monopoly or duopoly or ... comes via business relevance - not by standardization. Here also I am lost - are we saying that all the hoopla about standardization is against a known or unknown cloud provider ?
== "Cloud computing is here today, but we are very, very early"
Agreed. Of course, this clashes with other statements where it says we already have a monopoly
== "We should not waste time having an official cloud computing definition of 'interoperable.'"
Agreed. Instead, maybe we should define "interchangeable" and move on o;) (Sorry, couldn't resist) Again, this is right - the domain is very new that we really do not know what it is. Me thinks, we should not smother it with too much standardization. What says thee ?
Hi Ricky. I'm working with around 10 companies who are all at
different stages of using EC2 right now. Around half of them are large
multi-national organisations whom most people will have heard of. All
of the use cases are interesting and the challenges face by each are
interesting.
One use case is for a large mobile Telco who are putting out two new
services on the cloud in late March. Each service has the application
hosted on EC2 but the data hosted internally. The reason they feel
they can do this now is that they have just come off a longish term
project to implement SOA / services internally. They now have a secure
services interface which is what is used to communicate with the
internal applications and pulls / pushed data needed / given. All of
this is done over SSL.
Another large media company is going to use the cloud to host an
application with data in the cloud solely. I can't give too much away
bit this a unique new type of TV show that uses the ability of getting
larges cale compute resources on the cloud to crunch some fairly large
amounts of data in real-time for the show. The TV show will be sold to
networks all over the world. This would not be practical to do from a
cost perspective if they had to use something else other than the
cloud.
> IMO, the "definition" of a buzzword is not very interesting and usually it is overloaded and can mean different things to different people. Look at what "characteristics" that the cloud brings may be more interesting. And there are 2 different points of view: the user and the provider.
> User's concern:
> ============
> - Total cost of ownership
> - Speed to deploy/undeploy infrastructure resources to meet changing business workload
> - Data sensitivity protection and audit compliance
> - Reliability and QoS guarantee
> Provider's concern:
> ===============
> - Efficient use of underlying infrastructure, usually achieved by physical sharing with logical isolation between tenants ("Virtualization" is a common means to achieve that).
> - Efficient resource scheduling tool that can handle dynamic regrouping of resources based on temporary usage patterns.
> - Efficient management tools that can keep track of the health, ease the debugging in a highly distributed and virtualized environment.
> Note that I haven't said whether the provider is the in-house IT or a public cloud provider because I don't think this is an important distinction. I also believe "enterprise cloud" will be competing with "public cloud" for a while. And a public cloud will always have the advantage of economy of scale and operation expertise while an enterprise cloud will always have the advantage when data sensitivity is a concern.
> I also think reality of cloud computing for large enterprise is always a hybrid environment. It will be very useful to get some guidelines from this expert group on what should be put on the public cloud versus enterprise cloud, and how does components living in different domains can communicate with each other in a secure way.
> Rgds,
> Ricky
> ----- Original Message ----
> From: Frank D. Greco <fgr...@javasig.com>
> To: cloud-computing@googlegroups.com
> Sent: Sunday, January 25, 2009 2:10:47 PM
> Subject: [ Cloud Computing ] Re: Is "Internal Cloud" anoxymoron?
> At 05:10 PM 1/24/2009, natisha...@gmail.com wrote:
> >In my view what will see in the near future is hybrid-cloud i.e. local-
> >IT+ Internet Cloud. A typical scenario would be to run testing on
> >public cloud and production internally. There are many other scenarios
> >that jutifies this type of hibrid model.
> Totally agree Nati. In the long-term *perhaps* production systems for
> large enterprises will move to an external cloud. In the near/middle
> term, test/dev environments are ripe for external clouds as long as
> the proper precautions about data security are addressed.
> At a large investment bank (remember those?) several years ago,
> I was responsible for moving components of a fixed income analytics
> system to an external data center run by a large well-known
> systems company. This was done to address over-capacity usage.
> Effectively it was a primitive "inside-the-firewall cloud", ie,
> low-ceiling clouds... ;)
> Much of the focus was on: reducing the network/data latency
> as much as possible, software licenses for the cloud machines
> and last, but definitely not least, was data security.
> My former employer was paranoid about data security for good reason.
> I'm sure other large corporations would have the same stance.
> Part of our agreements with our external customers mandated the
> data would never be "co-mingled" (actual lawyer verbiage in our
> contract) with other customers on the same physical machine. This
> defeats much of the economies of scale advantages of the
> IaaS provider.
> Cloud providers should probably avoid this type of agreement
> in their contracts.
> In the end, we had our external site working fine. However, the costs
> were *very* high. Network plumbing, data security products,
> renegotiating software licenses, telecom provider costs, replication
> of data to the external site (to minimize latency) and not
> to mention the additional cost of the infrastructure-service provider.
> Just some anecdotal info to add to the discussion...
> >The real question about private cloud is whether local-IT would ever
> >be able to compete with the cost and level of service of those of the
> >public cloud.
> Many corporations believe that their spin on IT is what differentiates
> them from their competition. One can argue that this philosophy
> was the driving force behind the creation of cloud services in the
> first place. Would Google and Amazon have used an IaaS service?
> It was their belief that they could do it better than anyone.
> I first heard about utility (ie, cloud) computing from Bill Joy
> in the late 90's. Even Bill said some large companies will continue
> to use their IT as a competitive weapon. But he did say a large
> segment of businesses would leverage utility computing. Bill Joy
> is rarely wrong... ok... the invention of the c-shell wasn't
> exactly that great... ;+
> >This is where the economy of scales plays an important role (See James
> >Liddle recent blog: The Economics of Cloud
> >http://vehera.jsn-server7.com/LiddleBlog/?p=234 > >) i.e. the bigger the data-center is the higher the chance the cost
> >and level of service will get better then those of smaller data-
> >centers.
> Having a very large data center means you need another large
> data center
> as a DR site. It may be better to have families of data
> centers working
> in a federation. It will be interesting to see how well Sun
> Microsystems
> manages their Q-Layer acquisition and virtual data-centers notion.
> We can draw analogies to the power-grid, but at a certain point, this
> analogy doesn't apply. Sort of like the common "a GUI is just like
> a car dashboard" and "building software is like building a house"...
> yeah it is... but yeah it isn't...
> > It may very well that over time public cloud will become the
> >only option just like energy, however its going to take a while till
> >this will happen if at all.
> Totally agreed Nati. I'm not 100% convinced it will be the
> only option
> for all companies in the long run.
> The difference between cloud-based IaaS and the energy
> grid is software. Software innovation has a *long* way to
> go before its
> mature like our energy infrastructure. Besides, large
> public cloud utilities
> eventually will act like internal IT when many customers
> share the same
> virtual/physical infrastructure. :)
Hi Ricky. I'm working with around 10 companies who are all at
different stages of using EC2 right now. Around half of them are large
multi-national organisations whom most people will have heard of. All
of the use cases are interesting and the challenges face by each are
interesting.
One use case is for a large mobile Telco who are putting out two new
services on the cloud in late March. Each service has the application
hosted on EC2 but the data hosted internally. The reason they feel
they can do this now is that they have just come off a longish term
project to implement SOA / services internally. They now have a secure
services interface which is what is used to communicate with the
internal applications and pulls / pushed data needed / given. All of
this is done over SSL.
Another large media company is going to use the cloud to host an
application with data in the cloud solely. I can't give too much away
bit this a unique new type of TV show that uses the ability of getting
larges cale compute resources on the cloud to crunch some fairly large
amounts of data in real-time for the show. The TV show will be sold to
networks all over the world. This would not be practical to do from a
cost perspective if they had to use something else other than the
cloud.
> IMO, the "definition" of a buzzword is not very interesting and usually it is overloaded and can mean different things to different people. Look at what "characteristics" that the cloud brings may be more interesting. And there are 2 different points of view: the user and the provider.
> User's concern:
> ============
> - Total cost of ownership
> - Speed to deploy/undeploy infrastructure resources to meet changing business workload
> - Data sensitivity protection and audit compliance
> - Reliability and QoS guarantee
> Provider's concern:
> ===============
> - Efficient use of underlying infrastructure, usually achieved by physical sharing with logical isolation between tenants ("Virtualization" is a common means to achieve that).
> - Efficient resource scheduling tool that can handle dynamic regrouping of resources based on temporary usage patterns.
> - Efficient management tools that can keep track of the health, ease the debugging in a highly distributed and virtualized environment.
> Note that I haven't said whether the provider is the in-house IT or a public cloud provider because I don't think this is an important distinction. I also believe "enterprise cloud" will be competing with "public cloud" for a while. And a public cloud will always have the advantage of economy of scale and operation expertise while an enterprise cloud will always have the advantage when data sensitivity is a concern.
> I also think reality of cloud computing for large enterprise is always a hybrid environment. It will be very useful to get some guidelines from this expert group on what should be put on the public cloud versus enterprise cloud, and how does components living in different domains can communicate with each other in a secure way.
> Rgds,
> Ricky
> ----- Original Message ----
> From: Frank D. Greco <fgr...@javasig.com>
> To: cloud-computing@googlegroups.com
> Sent: Sunday, January 25, 2009 2:10:47 PM
> Subject: [ Cloud Computing ] Re: Is "Internal Cloud" anoxymoron?
> At 05:10 PM 1/24/2009, natisha...@gmail.com wrote:
> >In my view what will see in the near future is hybrid-cloud i.e. local-
> >IT+ Internet Cloud. A typical scenario would be to run testing on
> >public cloud and production internally. There are many other scenarios
> >that jutifies this type of hibrid model.
> Totally agree Nati. In the long-term *perhaps* production systems for
> large enterprises will move to an external cloud. In the near/middle
> term, test/dev environments are ripe for external clouds as long as
> the proper precautions about data security are addressed.
> At a large investment bank (remember those?) several years ago,
> I was responsible for moving components of a fixed income analytics
> system to an external data center run by a large well-known
> systems company. This was done to address over-capacity usage.
> Effectively it was a primitive "inside-the-firewall cloud", ie,
> low-ceiling clouds... ;)
> Much of the focus was on: reducing the network/data latency
> as much as possible, software licenses for the cloud machines
> and last, but definitely not least, was data security.
> My former employer was paranoid about data security for good reason.
> I'm sure other large corporations would have the same stance.
> Part of our agreements with our external customers mandated the
> data would never be "co-mingled" (actual lawyer verbiage in our
> contract) with other customers on the same physical machine. This
> defeats much of the economies of scale advantages of the
> IaaS provider.
> Cloud providers should probably avoid this type of agreement
> in their contracts.
> In the end, we had our external site working fine. However, the costs
> were *very* high. Network plumbing, data security products,
> renegotiating software licenses, telecom provider costs, replication
> of data to the external site (to minimize latency) and not
> to mention the additional cost of the infrastructure-service provider.
> Just some anecdotal info to add to the discussion...
> >The real question about private cloud is whether local-IT would ever
> >be able to compete with the cost and level of service of those of the
> >public cloud.
> Many corporations believe that their spin on IT is what differentiates
> them from their competition. One can argue that this philosophy
> was the driving force behind the creation of cloud services in the
> first place. Would Google and Amazon have used an IaaS service?
> It was their belief that they could do it better than anyone.
> I first heard about utility (ie, cloud) computing from Bill Joy
> in the late 90's. Even Bill said some large companies will continue
> to use their IT as a competitive weapon. But he did say a large
> segment of businesses would leverage utility computing. Bill Joy
> is rarely wrong... ok... the invention of the c-shell wasn't
> exactly that great... ;+
> >This is where the economy of scales plays an important role (See James
> >Liddle recent blog: The Economics of Cloud
> >http://vehera.jsn-server7.com/LiddleBlog/?p=234 > >) i.e. the bigger the data-center is the higher the chance the cost
> >and level of service will get better then those of smaller data-
> >centers.
> Having a very large data center means you need another large
> data center
> as a DR site. It may be better to have families of data
> centers working
> in a federation. It will be interesting to see how well Sun
> Microsystems
> manages their Q-Layer acquisition and virtual data-centers notion.
> We can draw analogies to the power-grid, but at a certain point, this
> analogy doesn't apply. Sort of like the common "a GUI is just like
> a car dashboard" and "building software is like building a house"...
> yeah it is... but yeah it isn't...
> > It may very well that over time public cloud will become the
> >only option just like energy, however its going to take a while till
> >this will happen if at all.
> Totally agreed Nati. I'm not 100% convinced it will be the
> only option
> for all companies in the long run.
> The difference between cloud-based IaaS and the energy
> grid is software. Software innovation has a *long* way to
> go before its
> mature like our energy infrastructure. Besides, large
> public cloud utilities
> eventually will act like internal IT when many customers
> share the same
> virtual/physical infrastructure. :)
On Saturday 2009.01.24, at 14:10 , natisha...@gmail.com wrote: [...]
> Based on my experience there are two main driving force:
> 1. On demand computing i.e. the ability to get a resource when i need > it in matters of minutes. > 2. Pay - per use i.e. the ability to pay only for what i use.
I quite like this characterization precisely because it's clean and simple.
I think a lot of the arguments to date end up being various forms of echo chamber / navel gazing amongst the insiders rather than the consumers of the services.
> The rest is an implementation details. > From an end user perspective if it doesn't matter that much if i can > get the above value from internal or external cloud.
As Frank noted, it can matter a great deal for various factors such as security. That said, I'd expand that into the general category of factors (and people who) need (to feel that they need) to be in control (whether that's an internal need or an externally mandated need).
> In my view what will see in the near future is hybrid-cloud i.e. > local- > IT+ Internet Cloud.
I totally agree. I see it as a spectrum rather than an either/or. Some people/projects need very little in-house while others need more.
> A typical scenario would be to run testing on > public cloud and production internally. There are many other scenarios > that jutifies this type of hibrid model.
> The real question about private cloud is whether local-IT would ever > be able to compete with the cost and level of service of those of the > public cloud.
If you add in the fully-burdened costs of things like data transfer in/ out, sensitive data security (and the attendant liability risk), need for some places of very high performance (e.g., predictable / low latency, large bandwidth, etc.), etc. then the answer is/can be yes. The variability for different organizations / projects needs is why it's a spectrum rather than competing point solutions.
> This is where the economy of scales plays an important role (See James > Liddle recent blog: The Economics of Cloud http://vehera.jsn-server7.com/LiddleBlog/?p=234 > ) i.e. the bigger the data-center is the higher the chance the cost > and level of service will get better then those of smaller data- > centers. It may very well that over time public cloud will become the > only option just like energy, however its going to take a while till > this will happen if at all.
I don't see that happening until it's far enough out to be science fiction. :-) There's too many specialized constraints, for one.
On Sunday 2009.01.25, at 09:57 , Stuart Charlton wrote: [...]
> On Sun, Jan 25, 2009 at 4:25 AM, Jan Klincewicz <jan.klincew...@gmail.com > > wrote: > It is a good idea for test & dev servers to be put into use to > accomodate seasonal fluctuations. However, in a modern, highly > virtualized shop, these machines no longer represent a lot of spare > capacity.
> - Most IT shops are not highly virtualized yet, even the "modern" > ones. This is changing, of course....
> - I've observed many a production outage because the organization > could not allocate the resources to perform an adequate stress test > of major changes, and the budget to acquire such an environment *for > the purposes of that project* would have been prohibitive.
I've run into this a *lot*. This is one of those problems that's so prevalent that people just seem to accept it as being part of the natural order of things (and that's scary).
One of my biggest rants is how these sorts of constraints affects how people (don't) see the problems that have and what the possible solutions are and how that perverts their thinking and decisions. I see this across the organizational layers and it's shockingly consistent (in a bad way). One of my hopes of the benefits of the hype of CC (regardless of the long-term reality) is that it helps to force open people's brains as to there being a lot of options that don't fit their existing biases.
> - I have seen two poles of behaviours with IT organizations related > to dev/test: either dev/test is ignored and teams have to scavenge > for resources, or huge amounts of idle capacity are tagged as dev/ > test. Both have major consequences, though usually the former is > much worse.
Indeed!
This is an area where I see a lot of value in the use of virtualization. Separating things logically and so being able to do dev/test on much smaller number of physical boxes than production while still getting lots of value from the logical testing -- this can actually be an advantage over perfect replicas of the production site since we can easily induce lots of the edge cases.
[...]
> I don't expect a whole sale jump into any of the above, but I do > expect some early adoptors getting into the act if they can see a > positive NPV with a (say) 12-24 month window....
Indeed.
Though I do see more and more startups (particularly web startups) jumping into this whole hog from day one.
"As Frank noted, it can matter a great deal for various factors such as
security. That said, I'd expand that into the general category of factors
(and people who) need (to feel that they need) to be in control (whether
that's an internal need or an externally mandated need)."
Yep I think you summarized it well.
Note that when I mentioned that from an end user perspective it doesn't
matter that much who is the service provider I'm assuming that there would
be other cloud providers that will provide the type of SLA's that are
required by most data-centers. Obviously that statement doesn't fit them all
and I'm sure that those who already maintain a large data-center would
still maintain large part of their operations in the private cloud. We
should note that organizations with large data-centers are probably the
minority not the majority.
[mailto:cloud-computing@googlegroups.com] On Behalf Of John D. Mitchell
Sent: Wednesday, January 28, 2009 8:49 PM
To: cloud-computing@googlegroups.com
Subject: [ Cloud Computing ] Re: Is "Internal Cloud" an oxymoron ?
Hi Nati,
On Saturday 2009.01.24, at 14:10 , natisha...@gmail.com wrote:
[...]
> Based on my experience there are two main driving force:
> 1. On demand computing i.e. the ability to get a resource when i need
> it in matters of minutes.
> 2. Pay - per use i.e. the ability to pay only for what i use.
I quite like this characterization precisely because it's clean and
simple.
I think a lot of the arguments to date end up being various forms of
echo chamber / navel gazing amongst the insiders rather than the
consumers of the services.
> The rest is an implementation details.
> From an end user perspective if it doesn't matter that much if i can
> get the above value from internal or external cloud.
As Frank noted, it can matter a great deal for various factors such as
security. That said, I'd expand that into the general category of
factors (and people who) need (to feel that they need) to be in
control (whether that's an internal need or an externally mandated
need).
> In my view what will see in the near future is hybrid-cloud i.e.
> local-
> IT+ Internet Cloud.
I totally agree. I see it as a spectrum rather than an either/or. Some people/projects need very little in-house while others need more.
> A typical scenario would be to run testing on
> public cloud and production internally. There are many other scenarios
> that jutifies this type of hibrid model.
> The real question about private cloud is whether local-IT would ever
> be able to compete with the cost and level of service of those of the
> public cloud.
If you add in the fully-burdened costs of things like data transfer in/ out, sensitive data security (and the attendant liability risk), need
for some places of very high performance (e.g., predictable / low
latency, large bandwidth, etc.), etc. then the answer is/can be yes. The variability for different organizations / projects needs is why
it's a spectrum rather than competing point solutions.
> This is where the economy of scales plays an important role (See James
> Liddle recent blog: The Economics of Cloud
http://vehera.jsn-server7.com/LiddleBlog/?p=234 > ) i.e. the bigger the data-center is the higher the chance the cost
> and level of service will get better then those of smaller data-
> centers. It may very well that over time public cloud will become the
> only option just like energy, however its going to take a while till
> this will happen if at all.
I don't see that happening until it's far enough out to be science
fiction. :-) There's too many specialized constraints, for one.
|
