Are Chargeback Systems needed beforer Wider Cloud Adoption?

[Ronnie Ray]

IT infrastructure purchases in  enterprises are frequently tied to business unit budgets that fund them. As long this infrastructure was physical, tying budgets to asset via asset management systems or CMDB’s was still possible.

 

With the cloud’s dynamicity and pooled use of resources how are enterprises going to manage cost and budget allocation to specific business units and processes? Are there monitoring and metering systems being launched that address this issue? Is this holding up cloud adoption in large enterprises that need a level of maturity in Chargeback systems before they bring the cloud in or scale out?

 

 

Ronnie Ray

www.marketplane.net/blog

[Paulo Calcada]

Hello Ronnie,

This subject is also discussed here,

http://groups.google.ca/group/cloud-computing/browse_thread/thread/9437e5f161ec0ab8

And, in my perspective, the solution should be based on the work done with to manage the budget available to communications (traditional voice communications)

Paulo

2009/5/5 Ronnie Ray <ron...@marketplane.net>

[Nik Simpson]

For many organizations, lack of charge back is a symptom of a wider
problem; the inability to accurately model costs for IT. If you can't
say how much a particular application costs, it's hard to create a
usable charge back model.

Funnily enough, I'm about to start research on data center cost models
if anybody is interested in having that discussion.

--
Nik Simpson

[Miha Ahronovitz]

----quote----

For many organizations, lack of charge back is a symptom of a wider
problem; the inability to accurately model costs for IT. If you can't
say how much a particular application costs, it's hard to create a
usable charge back model.

---------end quote--------

I think there is case when legacy accounting models are applied to a new technology.
What cloud computer does it eliminates the need for detailed business models.

The application and infrastructure should have an utility cost, that is monitored and charged.
That is how much it cost to run that application. These charges back could be accumulated to a credit card for the project / department.
If the credit card reaches a certain limit, perhaps a warning is sent to the project owner

The expense side of the cloud owners inside the same enterprise is a different point of profit and loss.
They operate a company credit card to monitor and pay for the AWS bills.

Unfortunately it is very hard to correlate the corporate credit card charges with the project department credit card charges.
But one can live very well without this correlation.
Simply if the corporate costs are increasing, the unit prices for internal projects are adjusted.

Miha

[Rao Dronamraju]

"I'm about to start research on data center cost models
if anybody is interested in having that discussion."

I am very interested in that discussion....

I am interested knowing whether any company out there does what we do with
alignment of business processes with IT solutions.

Are there any models, solutions out there that map/align corporate revenues
and expenditure with IT revenues and expenditure?....

I guess it would be very difficult to quantify revenues of an IT
organization w.r.t the revenues of a company....How can you say that this
portion of total revenues are because of IT?...same problem with probably
groups like HR....support groups....

Although you can know the expenditure of an IT organization, because you
cannot quantify the revenues, you cannot say an IT organization is a
profitable/non-profitable entity.

[Nik Simpson]

On 5/5/2009 1:49 PM, Miha Ahronovitz wrote:
>
> ----quote----
> For many organizations, lack of charge back is a symptom of a wider
> problem; the inability to accurately model costs for IT. If you can't
> say how much a particular application costs, it's hard to create a
> usable charge back model.
> ---------end quote--------
>
> I think there is case when legacy accounting models are applied to a new technology.
> What cloud computer does it eliminates the need for detailed business models.

Yeah right, I'd love to try pitching "What cloud computer does it
eliminates the need for detailed business models." to a VC.

>
> The application and infrastructure should have an utility cost, that is monitored and charged.

But for an enterprise data center the question is whether the cloud will
be cheaper as a solution overall for a particular workload, and thats a
very difficult question for most data centers to answer because they
don't have a good model for calculating their own costs.

> That is how much it cost to run that application. These charges back could be accumulated to a credit card for the project / department.
> If the credit card reaches a certain limit, perhaps a warning is sent to the project owner
>
> The expense side of the cloud owners inside the same enterprise is a different point of profit and loss.
> They operate a company credit card to monitor and pay for the AWS bills.

The point is that most enterprise data centers couldn't tell you with
any accuracy what a particular application costs to run.

It's the cost model for the existing data center that I'm researching,
not the cost model for a cloud application.
--
Nik Simpson

[ep...@weiglewilczek.com]

<quote>

Although you can know the expenditure of an IT organization, because
you
cannot quantify the revenues, you cannot say an IT organization is a
profitable/non-profitable entity.

</quote>

from my experience the expenditure part isn't easy either. ask two
companies what exactly is contained in their it costs/budget and you
will probably get two very different answers.


On 5 Mai, 19:45, "Rao Dronamraju" <rao.dronamr...@sbcglobal.net>
wrote:

[ART PERERA]

Having worked for a brilliant ASP some time ago, I figured out how they were so successful.
Hire a good cost accountant with a ASP or ISP background that can calculate all the cost of the nuts and bolts that it takes to provide a Cloud service.
Once you've covered all the items, you then know what to charge and what profit margin you are
going to make. It sounds simple, but there are so many things that are overlooked. And if you look at
all the failed businesses, you will note that they did not know at what point they were making a profit or
losing money.


Most businesses fail at performing this simple task up front. If you don't do this first, you will fail.

[Miha Ahronovitz]

Nik, nice conversation

> It's the cost model for the existing data center that I'm researching...

This is exactly the reason you have difficulties.

> Yeah right, I'd love to try pitching "What cloud computer does
> it eliminates the need for detailed business models."

I should clarify. What we no longer need is to agonize to death looking for the a generalized cost business model for the existing data center. It doesn't exist, all we have local practices, good or bad. Gartner came out with things like TCO, that no two people calculate the same answer... I assume you want black on white a proof that the cloud computer will lower the cost per application compared to what we do now.

We don't know now, as an instant answer, - in any data center - whether the data center makes money or looses money. If the CEO calls the CIO and tells him: please lay off 15% of your staff, the CIO executes because, everyone must do what the CEO asks and make sacrifices.

But if the company operates a cloud, it knows how much it pays out to AWS-like services, how much are the operating costs, how much they charged for the services cotaining the applications, It knows instantly how much momney the cloud makes. So when the CEO calls with demand for laying of 15%, the CIO says: sorry, we are making $1M a month (or a year), please cut somewhere else, here are the data, INSTANTLY.

What is more expensive: $1M to buy a private plane to enjoy myself, or an investment of $2M to make a profit of $1M? It is not how much it costs, but how much money I make. This is what the VC wants: profits, not cost reductions by themselves in isolation, which can mean nothing.

The secret is in the built in accounting-billing that must be done in a cloud by definition (gives the illusion of infinite resources and I pay only for what I use)

Cheers,

Miha

----- Original Message ----
From: Nik Simpson <n...@alaweb.com>
To: cloud-c...@googlegroups.com
Sent: Tuesday, May 5, 2009 12:29:19 PM
Subject: [ Cloud Computing ] Re: Are Chargeback Systems needed beforer Wider Cloud Adoption?

[Jim Starkey]

Miha Ahronovitz wrote:
> Nik, nice conversation
>
>
>> It's the cost model for the existing data center that I'm researching...
>>
>
> This is exactly the reason you have difficulties.
>
>
>> Yeah right, I'd love to try pitching "What cloud computer does
>> it eliminates the need for detailed business models."
>>
>
> I should clarify. What we no longer need is to agonize to death looking for the a generalized cost business model for the existing data center. It doesn't exist, all we have local practices, good or bad. Gartner came out with things like TCO, that no two people calculate the same answer... I assume you want black on white a proof that the cloud computer will lower the cost per application compared to what we do now.
>
> We don't know now, as an instant answer, - in any data center - whether the data center makes money or looses money. If the CEO calls the CIO and tells him: please lay off 15% of your staff, the CIO executes because, everyone must do what the CEO asks and make sacrifices.
>
> But if the company operates a cloud, it knows how much it pays out to AWS-like services, how much are the operating costs, how much they charged for the services cotaining the applications, It knows instantly how much momney the cloud makes. So when the CEO calls with demand for laying of 15%, the CIO says: sorry, we are making $1M a month (or a year), please cut somewhere else, here are the data, INSTANTLY.
>
> What is more expensive: $1M to buy a private plane to enjoy myself, or an investment of $2M to make a profit of $1M? It is not how much it costs, but how much money I make. This is what the VC wants: profits, not cost reductions by themselves in isolation, which can mean nothing.
>

If it's significantly more expensive to do something in the cloud, you
argument falls apart.

Apples to apples, in-house almost always wins. If you're going to
justify going outside, you have demonstrate a value proposition.

VCs, I believe, have largely learned the difference between hype and
reality. And, Miha, since you have never been interested in actual
costs, your arguments lack substance.

There is value, great potential value, in the cloud. But it isn't a
simple as paying Amazon and shutting down existing, paid for servers.
And that value has to be something that the cloud can do better -- not
just more cheaply -- than you can do in house. That is what is going to
excite the VCs.

> The secret is in the built in accounting-billing that must be done in a cloud by definition (gives the illusion of infinite resources and I pay only for what I use)
>

Illusions crash, and crash hard. If nobody is willing to put their
signature on the legal line that the capacity is there for the asking,
no rational person is going to count on it. "Come, and they will have
built it" works, but only in fantasy.

[JL Valente]

Miha,

We are again talking about enterprise IT.
Enterprises want cost predictability and transparency so that they know
what and when to throttle up and down.
In an enterprise IT world that is becoming more virtualized every day
and shaped as a shared service organization across networking, storage
and compute, the allocation of IT costs to business units becomes much
more challenging.

There is also little "pay per use" today and even the pay per use in
existence is not for what you use but what you order (like in the
restaurant and I have not seen a doggy bag for IT leftover yet). IT
costs cover many areas beyond networking (15%+ of IT budget), computing,
software licensing (mostly perpetual), facilities and hardware. Just
think of the people costs.
Likewise pay per use without policies and controls is the best recipe
for cost runaway, just like when you use your data plan on a mobile
phone in a different country.

In our current economic climate, cost reduction is a mandate across the
board so enterprise IT must understand the cost equation well enough,
set policies and controls so that they can help business units and
executive management decide where to invest, reallocate, eradicate or
outsource

-JL

[Ray Nugent]

I agree. It has to be about value, of which cost is one possible variable. A cost only proposition is going to be hard to sell unless you add an accountant to your sales team.

Ray

---

From: Jim Starkey <jsta...@NimbusDB.com>
To: cloud-c...@googlegroups.com
Sent: Tuesday, May 5, 2009 4:34:06 PM

[Miha Ahronovitz]

Jim, this is a splendid post. Thank you.

You are right. The Luddites, will procrastinate to ask for a comparison of cost now and and in the cloud. No one on this thread knows a standard way how to calculate the costs of delivering an application inside a data center today. So we will postpone for ever the cloud adoption. Same way immigration department asks for papers

For example if I open a bakery, the first loaf of bread costs $10,000, the second $5,000 and the 10,000 th loaf costs me 2 cents

> Apples to apples, in-house almost always wins.

Sure. The private cloud is an in-house implementation. But this is case for study.. Which cloud to select? However the cloud solution, once optimized, is usually the best. This is why.

Here I have a bunch of computers. (1) If I operate them as a data center, I have a challenge to see who is using what and a challenge to justify the costs. The costs may be variable. One project may use an application 24 hours a day in a week and nothing for 10 weeks after that. Which cost per day I take? How do I predict the future usage?

(2) If I operate the same bunch of computers as private cloud, assuming I have the right software and services associated for elasticity on demand, I know exactly what I billed to whom and when and why and how much I spent..

Isn't (2) cleaner than (1)? Sure we can assume some costs to do calculations. Let's use the bakery example. I establish the price of $2.00 per day to deliver one loaf of bread to every subscriber, I will loose money if I have only 2 users, but if I have 10,000 users My margin is 1,98 /2,00 = 99%

Have a wonderful evening,

[Miha Ahronovitz]

JL, it makes a lot of sense what you aptly describe.
On each individual case, sure a calculation of costs can be made and must be made.
What Nik argued, is that this generic business model for cost calculation per application is difficult to create.

The word "costs" on this thread shows our legacy preoccupation with costs, not profits
What cloud computing (including private cloud) offers, is a fresh look at we call now data center.

No one gets rich by reducing costs only. One gets rich by making a profit. I think an efficient IT operation inside an Enterprise has the right to make a profit locally, reflecting it's efficiency.This requires a mind shift.

[William Louth (JINSPIRED.COM)]

Education of users in terms of their usage of software and indirectly system resources is a important element of charge back schemes which have typically failed when applied based on metrics that have very little relation to the actual usage patterns by users/profiles/departments/units. It is for this reason that most management accountants use activity based costing. I think we also miss the point of charges. They have a relation to cost but they are not necessarily equal if so then how could the IT management teams get funding for additional service improvements. IT management need to add an element of profit into charges to support their own internal initiatives which will eventually drive down costs and charges.

http://en.wikipedia.org/wiki/Activity-based_costing

ABCs of Activity Based Management: Crushing Competition Through Performance Improvement
http://books.google.com/books?id=lUwq_NwkCIEC&printsec=frontcover

A Unified Approach to Performance Management and Cost Management for Cloud Computing
http://www.jinspired.com/products/jxinsight/meteringthecloud.html

William

[Bernard Golden]

To many of the points in this thread:

1. Most IT orgs have very limited insight into direct cost of
providing service. Very few implement any kind of granular chargeback.
I applaud the Jinspired initiative toward ABC. I addressed this issue
in a recent blog post:

http://www.cio.com/article/480595/The_Case_Against_Cloud_Computing_Part_Four?taxonomyId=168354

2. On IT cost vs. profit. Most companies look askance at internal
groups attempting to charge above their costs to other internal
departments. So I think it's unlikely that cloud computing (or any
other technology initiative) is going to change the way companies work
internally.

3. With regard to the assertion that cost reduction isn't a successful
business model, and therefore cloud computing can't be successful with
a cost focus, that's an overstatement (or, put more bluntly, just
plain wrong). One has only to look at the outsource manufacturing
companies like Flextronics to see an industry built on cost reduction.

4. The question of how important chargeback will be in private cloud
offerings is uncertain. VMware will release a chargeback module as
part of its CloudOS later in the year, which suggests it's important
and, by virtue of VMware's prominent position in data centers, likely
to be an important part of private cloud conversations.

5. Setting aside the question of whether internal IT groups must, by
virtue of not attempting to make a profit, deliver computing
capability less expensively than for-profit external providers like
AWS, just the fact that the external providers can offer transparent
pricing certainty will be attractive to many business units that get a
coarse-grained IT chargeback with no transparency as to pricing
rationale. In my view, the move to chargeback transparency demand will
be akin to the pressure placed on IT groups a few years ago when end
users demanded internal systems as convenient to use as Internet apps
-- but with money behind the argument, it may be higher pressure!

[Pietrasanta, Mark]

Providing use and cost data for an in-house data center is pretty simple. As is providing actual and projected usage data. I'm not sure why you think this is so difficult?

You seem to want to justify cloud usage because you'll get a bill telling you what you used, versus you having to put some very simple things in place to measure your in-house usage? Seems very silly to me, but perhaps I'm missing something.

The fact is, there is no current model that makes the large commercial clouds even close to cost justified. And there's really no feature/function benefit at all - in fact, there's a *huge* downside in the "terms of use" and complete lack of accountability by the large cloud provider.

Small private clouds, aka IaaS, aka shared virtualized environments, can be extremely cost effective, depending on the pricing model and application architecture. These folks generally provide solid terms of use, strong SLAs, and everything else you'd expect from a hosting provider, but at a significantly reduced cost (with no reduction in security).

[mij...@sbcglobal.net]

Yes. Yes. And once more Yes!

The very nature of the cloud is to know at any time what additional resources are needed and why.
I know people look askance today. Cloud computing without a chargeback or accounting is like a car with no gasoline.

mihq
Sent from my Verizon Wireless BlackBerry

-----Original Message-----
From: Bernard Golden <bernard...@gmail.com>

Date: Wed, 6 May 2009 06:43:37
To: Cloud Computing<cloud-c...@googlegroups.com>

[Pietrasanta, Mark]

That makes no sense. To say someone will go with a large commercial cloud provider simply because they prefer a detailed bill, even though it's double or triple their internal costs?

Never.

As long as internal IT chargeback models are substantially cheaper than large commercial cloud pricing, it doesn't matter what detail the "bill" provides. Couple that with the complete lack of accountability in the large commercial cloud, and there is no decision to be made.

That all being said, internal IT shops are starting to look at how to provide more detailed accounting of costs and chargebacks, which is a nice side effect. Especially as they move towards a virtualized model where charging a project for a physical machine no longer makes much sense. It doesn't change any of the fundamentals of the decision making, however.

[Miha Ahronovitz]

----Quote---
...even though it's double or triple their internal costs?
---end quote----

You are attributing me words I never said. The question on the Subject line is:

"Are Chargeback Systems needed before WiderCloud Adoption?"

.. which means someone has decided to have at least a small private cloud, and wants to adopt it on a wider scale.

My answer is Yes, Yes, Yes. There is a must for a billing system.

Debate is good, and it will boring all of us to have the same opinions.

There is a majority group who claims we must know what an application cost is in a data center, before we decide the go to a cloud. We all agreed there is no generic way to calculate the internal costs. How do you know whether the cloud costs 2x or 3x, when x internal costs are not known in 90% of the cases?

If we know this cost, then yes, we establish a price for that application delivered as a service, and build a cloud capable to deliver competitively. Surely it does not make sense to put an application with 2 users in the cloud, unless, maybe, I can share it externally with another 1,000 users, and bring the price down.

Now if I use external users, I MUST have a billing system. The Enterprise CIO becomes an entrepreneur.

The moment a bill from AWS arrives, we face a chain of events without precedent in a legacy data center. How do we share this bill? How do we share everything? If we must deliver for highest ROI to shareholders, what about the budget of $100M dollars invested in a large data center , maybe yearly? Can we calculate the ROI? Can we bring external users as customers?

To many of DC executives, what I say is heresy: To charge internally for IT services? To make a profit? To get external users, when possible? In a Data Center?

Seneca said: "Nullum Magnum Ingenium sine mixtura dementiae fuit".
There is no great ingenious idea without a touch of madness

This is the real cloud computing.

Miha

[Greg Pfister]

"2. On IT cost vs. profit. Most companies look askance at internal
groups attempting to charge above their costs to other internal
departments."

From my experience: ROTFLMAO, because while literally true, it's often
irrelevant -- the "costs" are easily grossly inflated. IT departments
usually have captive audiences, and great incentives to increasing
their budgetary empire.

This is a major reason, by the way, that the "in house is always more
efficient" line doesn't get very far with me. I agree that it can be
more efficient. But internal IT is not in a competitive environment;
external clouds are.

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

On May 6, 8:43 am, Bernard Golden <bernard.gol...@gmail.com> wrote:
> To many of the points in this thread:
>
> 1. Most IT orgs have very limited insight into direct cost of
> providing service. Very few implement any kind of granular chargeback.
> I applaud the Jinspired initiative toward ABC. I addressed this issue
> in a recent blog post:
>

> http://www.cio.com/article/480595/The_Case_Against_Cloud_Computing_Pa...

[Pietrasanta, Mark]

Who agreed with that premise? I certainly think it's fairly easy to provide internal costs (within a range of accuracy, but good enough).

Further - If you move the application to the cloud, there should be a corresponding internal savings, right? If so, that savings better be more than the external costs (or the external functionality better justify the increased total costs), otherwise that application is coming back in house quickly.

Again, I think the premise is flawed. At least in every business I've been in, the internal data center/IT costs have always had to be justified, correlated to something - activities/applications/business units/etc. - and then a plan put in place to project costs/revenue.

Charge-back inside an organization is sometimes meaningless, as it ultimately all rolls up anyway. But cost allocation is almost always done, if for no other reason than to show where and how the money is being spent. And often some sort of pseudo-chargeback occurs, if for no other reason than to provide some friction to the business units so they don't treat IT as "free".

Managing and allocating data center/IT costs is nothing new, and nothing magical. So I assume I'm missing something in your premise?

[Rao Dronamraju]

Just curious, I know that many enterprises atleast have deployed SAP in the
last 10 years. Isn't SAP's FICO supposed to provide the financial and cost
center accounting and control for these enterprises and their
departments?...Won't the IT center/CIO's department one of the cost centers
that gets SAPed and FICOed?....so there are probably enterprise out there
that do account the costs of IT Data Centers pretty well?....is it not?....

But what you may not know is the cost per application or cost per solution
etc?....but you can know the overall expenditure of a data center...

[mij...@sbcglobal.net]

I smiled There is more than one truth in the world... maybe we are both right. The future will tell. Not all DC will become clouds. The majority will

:-)

M

Sent from my Verizon Wireless BlackBerry

-----Original Message-----

From: "Pietrasanta, Mark" <Mark.Pie...@aquilent.com>

Date: Wed, 6 May 2009 13:52:00
To: cloud-c...@googlegroups.com<cloud-c...@googlegroups.com>

[Pietrasanta, Mark]

You certainly do know the cost per "person paying" - you have to, or you don't get funded (at least everywhere I've been). Whether that's an application, or set of servers, or whatever, the granularity is down to the level of what's needed for the business unit buying the service. And of course there's the general overhead of IT/internal data centers, which becomes basically a tax to everyone.

It doesn't always need to be just the "application", since often the costs need to cover more than that. There are significant costs that impact this, such as SLA needed, training, software development, other support, etc.

The "cloud" addresses only a single portion of the costs - basically hosting/hardware - and frankly does it at a substantial premium. For what? Improved billing itemization? Not a compelling selling point.

So the granularity of costs for internal IT is based on what makes business sense. They can certainly be improved, and the cloud usage-based billing appears to be causing many IT departments to revisit how they allocate costs, or at least how they document it. This is great.

But it does little to provide any cost justification to large commercial clouds.

(HPC and highly dynamic academic use is something completely different, and usage-based computing is great. But again this is a billing model that's resulting in the benefit, that's it.)

[Nik Simpson]

On 5/6/2009 1:09 PM, Rao Dronamraju wrote:
>
>
> Just curious, I know that many enterprises atleast have deployed SAP in the
> last 10 years. Isn't SAP's FICO supposed to provide the financial and cost
> center accounting and control for these enterprises and their
> departments?...Won't the IT center/CIO's department one of the cost centers
> that gets SAPed and FICOed?....so there are probably enterprise out there
> that do account the costs of IT Data Centers pretty well?....is it not?....
>
> But what you may not know is the cost per application or cost per solution
> etc?....but you can know the overall expenditure of a data center...
>

Yes, but we need something a little more granular ;-)

--
Nik Simpson

[Rao Dronamraju]

You can have granular accounting. All the material costs (software, HW,
power (power just to the data center may not be accounted separately) etc
and people (salaries) costs can be accounted for. So you can get granular
costs...it is just that they are not aligned on a solution basis. If they
can align it on a solution basis, you can align the solutions to business
processes and get direct relationship between business processes and IT
expenditure...this would take a decade may be:-)

-----Original Message-----
From: cloud-c...@googlegroups.com
[mailto:cloud-c...@googlegroups.com] On Behalf Of Nik Simpson
Sent: Wednesday, May 06, 2009 1:59 PM
To: cloud-c...@googlegroups.com
Subject: [ Cloud Computing ] Re: Are Chargeback Systems needed beforer
WiderCloud Adoption?

[Pietrasanta, Mark]

If by that you mean that most data centers will leverage virtualization, either internally or by using a private "cloud" at a hosting facility, I agree.

However, if you mean most DC will end up in EC2 (and the like), I couldn't disagree more.

[William Louth (JINSPIRED.COM)]

This is handled within our metering model by supporting multiple named (cost) groups per metered activity with each name being a composite that represents a hierarchical cost (object) structure. You can view the metering of software activity and resource consumption from the perspective of lets say a user (company/dept/user), a web address (protocol/domain/path/...), the service (service/application/operation/task), the actual codebase (packages/class/method). This is easy within software. Trying to do this from bread crumbs left in a database tx trail is a completely different matter and one that leads to many failures in Business Insight or Business Activity monitoring solutions.

[Rao Dronamraju]

Seneca said: "Nullum Magnum Ingenium sine mixtura dementiae fuit".
There is no great ingenious idea without a touch of madness

Where the hell do you read all this CRAZY stuff!.

Yes, I agree with you very much.

Great Ingenious Thinking = Constant Madness Factor + (1 / Degree of Madness)
* Ingenious Thinking Ability...does that sound right?....

[Rao Dronamraju]

Interesting (brief) article on cloud business case?.....

The numbers don't lie
An independent study found on-site Microsoft apps - Office and Exchange -
cost 20x in capital dollars and 5x-6x more than Google Apps on a 3 year
Total Cost of Ownership (TCO) basis. How can Microsoft compete?

http://blogs.zdnet.com/storage/?p=472&tag=nl.e550

You may want to go to clarity and checkout the original slides....might need
to register...

http://clarity.twinstrata.com/reports.php

Regards,
Rao

[Pietrasanta, Mark]

This is different than CC, for a number of reasons:

1) Office is *already entrenched* in like 75% of enterprises. It's a sunk cost. So switching to Google Apps is an increase in cost for at least some period of time. (CC is the more expensive option but isn't the incumbent one, sort of the reverse.)

2) There are no terms of service/SLA to speak of for Google Apps. That means that only very small enterprises will ever use it, and/or the costs will skyrocket if you want a real SLA from Google.

3) The only reason Google does this is for access to the data and activities (in aggregate). Most enterprises will refuse to allow this, and will want a Terms of Use that specifically prevents this - which will never happen and/or come with a substantial increase in cost.

4) The change management is too big for most enterprises.

So other than small shops and local government, this is a non-starter until SLA/Terms of Use are improved (which will come with a corresponding cost increase, likely washing any ROI).

Office Live has a better chance.

-----Original Message-----
From: cloud-c...@googlegroups.com [mailto:cloud-c...@googlegroups.com] On Behalf Of Rao Dronamraju
Sent: Wednesday, May 06, 2009 8:35 PM
To: cloud-c...@googlegroups.com

[Bernard Golden]

Greg, you are of course right. Currently, hosting in most companies is
a monopoly provider situation, which is usually not conducive to cost-
effectiveness. My personal belief is that most IT organizations are
not competitive cost-wise with cloud providers, in part because
they've never been designed with a view to be fully automated,
homogenous infrastructure, etc., like the new cloud providers.

With respect to the comment made earlier in this thread responding to
my point that many business users (e.g., the sales department of a
company) would be attracted to cloud providers due to their cost
transparency (i.e., charges are directly tied to specific resource
use), someone stated that wouldn't happen when the cost of external
cloud providers are two to three times internal cost -- well, I guess
we will continue to have different opinions on this subject.

In my opinion, many IT users, when faced with a choice between a "it
will cost you $180/month" and "you will need to provide $5K to get
started on the project," will be attracted to the former, even in the
face of a rationale that, long-term, the latter alternative is less
expensive. Again, in my opinion, today these people have no choice --
they are offered the opaque lump sum option and nothing else, but in
the future, when another option that offers cost transparency is
available, it will be a different story.

On May 6, 10:39 am, Greg Pfister <greg.pfis...@gmail.com> wrote:
> "2. On IT cost vs. profit. Most companies look askance at internal
> groups attempting to charge above their costs to other internal
> departments."
>
> From my experience: ROTFLMAO, because while literally true, it's often
> irrelevant -- the "costs" are easily grossly inflated. IT departments
> usually have captive audiences, and great incentives to increasing
> their budgetary empire.
>
> This is a major reason, by the way, that the "in house is always more
> efficient" line doesn't get very far with me. I agree that it can be
> more efficient. But internal IT is not in a competitive environment;
> external clouds are.
>

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

[Pietrasanta, Mark]

Do you have any basis for that opinion?

In my *experience*, what you say is not consistent with actual activities. IT shops are constantly under pressure to reduce costs, and are flooded with examples of outsourced alternatives that seem to be less expensive.

IT shops are always performing analysis of what makes the most sense in the "big picture" for the enterprise, and having to provide justification for every dollar and headcount. For example, moving one application to the cloud could make integrating with that application much harder, or result in additional costs for connectivity/bandwidth/backups, etc. And we can ignore the terms of use/SLA for now.

I know my experience is not unique. I'm not dismissing your opinion, and I'd even say there are a number of old-school IT organizations out there, but I think that's a decreasing minority (at least here on the east coast - the foundation of my experience).

If other folks have different experience, I think that would be useful to hear.

[Greg Pfister]

"Do you have any basis for that opinion?"

Yes. Here's one of several cases: When I worked for IBM, I had a
chance to see some of the details of the internal chargeback for email
storage. The markups were utterly ludicrous - on the order of 1000X
the actual cost of storage. Certainly, managed storage is a far cry
from raw spinning stuff, but 1000X?

Just one data point, yes. And in somewhere like IBM, if you're really
competent with computing, do you aspire to work for its internal IT
department? I'm certainly willing to change my mind given other
specific examples.

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

On May 7, 9:52 am, "Pietrasanta, Mark" <Mark.Pietrasa...@aquilent.com>
wrote:

[Miha Ahronovitz]

Greg, I read with pleasure your points below.


1. IT departments usually have captive audiences, and great incentives to increasing their budgetary empire.
2. internal IT is not in a competitive environment; external clouds are.

Here we have the soon to be famous proof that the domain .gov saves 90% of infrastructure costs by going to a cloud.
In addition the portal is extremely secure

http://www.govtech.com/gt/654240

I read this:

"If GSA were a public company, our
$17.7 billion in revenues would place
us at 141 on the Fortune 500, ahead of
Google and Nike, among others."

http://www.gsa.gov/graphics/admin/Citizens_Report_011509_FINAL.pdf

The news of the US Government cloud portal, proved all past, current and future skeptics that yes, cloud computing saves money, big way and reduces complexity and handles peak traffic.. It makes money for everyone

If GSA services can do it, anyone can do it and go cloud.

It also shows where the $42B per year of IDC are. In companies like Terremark Worldwide, a Miami-based infrastructure services provider. Not anything revolutionary, just superb execution with existing technologies.

Miha

[Don Nalezyty]

Are chargeback systems needed before wider Cloud adoption?

Based on my experience, I believe the answer is yes.

I take Cloud in the broadest sense to mean all of is variants. I won't
open that can of worms with this group, but leaving aside XaaS, I
think Christopher Hoff has a good summary here:
http://www.rationalsurvivability.com/blog/?p=743 We've got several
cloud like entities internally, but expanding to leverage the benefits
of cloud via shared infrastructure, manageability and on demand usage
is currently hindered by our chargeback model or lack of one.

We don't have chargebacks for IT services managed by our internal IT
groups and our CIO doesn't want them. I think this is fairly common,
at least with those I've spoken to at other large enterprise
companies. There are exceptions as noted by Greg above, but I think
it is less common. It may even that examples like the one at IBM led
to a certain amount of backlash that seems to have put chargeback
models for IT services out of favor.

For us, the cost of IT is essentially a "tax" that goes to all of the
LOBs across the company. That "tax" is based on the total costs of
running the data center facilities, all of the IT headcount and
variables based on # of dedicated headcount in IT supporting the LOB,
# of HC using shared services and the majority occupier of a floor
tile within the data centers. In fact, we could get much more
granular today by tracking assets and their avg. power consumption
along with tracking of HC resources to support and administer
environments. Getting granular to the cost of VMs or even processing,
storage and network traffic is more difficult, but not unachievable.
Unfortunately, there's little desire to do this as it get's too close
to a chargeback model, which has already been sworn off as bad
business.

When you introduce virtualization to increase utilization, the most
benefit is derived by sharing resources in the largest pool possible.
The problem arises in how not having chargebacks impacts purchase of
the assets within the data center. In most cases, all HW including
servers, network equipment, storage, etc. are paid for by the LOB.
Without chargebacks there's no means for IT to purchase that HW and
charge the customer only for the portion they use. It's silly, but
there's just no way to allocate 3 tenths of a server each to 3
different LOBs.

This one item has been the single largest stumbling block to expanding
our internal clouds. No LOB is willing to pay for HW that is not
exclusively for their own use, so we're unable to put their resources
in a shared pool as part of a cloud. We've been able to build clouds
for some of the large LOBs that actually have enough servers to make
it effective. Even in those LOBs that have seen the value of cloud
computing are generally unwilling to allow us pool their resources
with other LOB's in a single larger cloud. Without buy-in of pooling
resources from those that are already closest to the model we're
proposing, no other LOBs seem to be willing to make the leap. There's
an additional catch-22 in the fact that IT can't just go and buy all
the HW, because only shared infrastructure like network distribution
or services like email that are leveraged by the entire corporation
get approval from executives.

This a cultural issue and requires a great leap of faith for people to
just say, "fine I'll pay for hardware to buy into the cloud, but from
that point I will no longer own those assets - I'm fine with a OLA or
SLA that says I have 'X' resources available to host my VMs."
Although there will still be FUD that will have some LOBs hold back,
chargebacks would simplify things and remove a significant barrier to
moving to the cloud.

So, that brings us around to other incarnations of cloud, such as
private external clouds; how can we convince executives this make
sense if we can't even convince LOBs it makes sense internally. There
needs to be an effective comparison of costs, and without chargebacks
it's very difficult to do an apples to apples comparison. It's
important that we have this level of comparison, because as has been
debated in this group often enough, it's unlikely any company is going
to move their entire IT infrastructure to an external cloud.

We've been working on a chargeback model and continue to refine it and
intend to use it for comparisons to help sell the c-level executives
on moving IT to a truly services based organization supporting those
services from cloud funded by chargebacks from all of the LOBs using
those services. Until we can make that huge cultural shift, we'll
just keep picking away at the model and show the numbers of how we
saved millions on the few LOBs that have moved to limited cloud
models.

Regards,
Don

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

> ...
>
> read more »

[Pietrasanta, Mark]

There's some irony to this, and your comments.  I was the primary technical person involved in this from their industry partner, and led much of the design and implementation of the solution.  The Gov't business owner is fortunately someone with a lot of vision, and she has the constant drive and determination to make things better.  Without her, I'm not sure this would have been successful - as she said, the issues were often *not* technical.  She needed to see where this could go, and be willing to "jump in", so to speak - and she did.

 

She's someone that I would put up against many leaders in industry.

 

But it's funny how people in this group quickly slant a very limited amount of data to try and support whatever claim they're trying to make. 

 

To help frame this up better - the Gov't moved from an external hosting provider and traditional hosting model to another hosting provider and a virtualized/*private* cloud model.  This did indeed result in substantial cost savings, and actually an *increase* in total capacity, as well as all the reliability/flexibility/etc. of moving to a private-cloud model.

 

I can assure you, this completely supports what I've been saying all along, including:

 

1) Large commercial clouds are completely cost prohibitive, and would have resulted in a cost *increase* to the Gov't.

 

2) The SLA (and terms of use) from the private cloud is enterprise-level, something unavailable from the large commercial clouds (today)

 

3) Security is at-or-higher than their previous implementation, again something difficult to achieve in the large commercial clouds

 

4) Private clouds actually *come with services*, so you can truly reduce the internal staff needed and/or raise the level of what your internal staff are doing (again, something not available from the large commercial clouds, today).

 

Private clouds are really the only thing that makes any sense right now, and are ultimately nothing more than a natural evolution of the hosting provider.  Unfortunately, many hosting providers seem to be incapable of making this step, which if evolutionary rules follow, means these providers will soon be extinct.

 

Had they been able to host internally using similar technology, it likely would have been even less expensive, but that simply isn't an option right now.

 

If you have any questions about this, I'd be happy to answer them (unless they fall into an area that I'm not allowed to).

-Mark

 

 

 

---

From: cloud-c...@googlegroups.com [cloud-c...@googlegroups.com] On Behalf Of Miha Ahronovitz [mij...@sbcglobal.net]
Sent: Thursday, May 07, 2009 6:03 PM

[Rao Dronamraju]

 

“To help frame this up better - the Gov't moved from an external hosting provider and traditional hosting model to another hosting provider and a virtualized/*private* cloud model.  This did indeed result in substantial cost savings, and actually an *increase* in total capacity, as well as all the reliability/flexibility/etc. of moving to a private-cloud model.”

 

Just curious, how did they achieve 90% savings by moving from a traditional hosting model to another hosting provider and a virtualized /private cloud model?...

What is in the private cloud model that contributed to such HUGE savings?....just virtualization?....

 

“1) Large commercial clouds are completely cost prohibitive, and would have resulted in a cost *increase* to the Gov't.”

 

I am in the camp that says we do not know yet whether large commercial clouds are viable are not?....No body has enough data to prove or disprove this at this time…

 

I think all of us in this group are debating an issues for which there just isn’t enough information either way…..

 

If you say large commercial clouds are not viable based security, compliance, governance etc….I can understand….but I am not sure about it based on cost factors....

 

I am also interested in learning their security, compliance, governance, visibility…etc…etc issues….they have been addresses primarily through SLAs?....any specific NEW technologies deployed?....I know this could be confidential…

 

Thanks

 

 

---

<BR

[Miha Ahronovitz]

Mark,

What do you mean by "a large commercial cloud", versus a private cloud?
I know the definition based on who has access to the cloud

- Internal only Private Clouds,
- Hybrid, seasonal or permanent Private Clouds
- External subscriptions Public clouds

Based on the definition above, I would classify GSA services as Hybrid Permanent Private cloud. And if it can handle RFPs or GSA-price-listed acquisitions with no competition, it also qualifies as a commercial cloud. Cloud Procurement on line saves and makes $ for the Government. and is is an indirect form of pay-per-use.

How a " large commercial cloud" is different from GSA's Private permanently hybrid cloud? And what is the definition of large (versus medium or small, I suppose).

Isn't the cloud lesson from GSA Services applicable universally, in essence?

Miha




--- On Thu, 5/7/09, Pietrasanta, Mark <Mark.Pie...@aquilent.com> wrote:

[justin.brister]

Mark,

you could not be more wrong;

1 - Companis making the move are realising approx. 300%+ ROI in the
first year, by using GApps to replace on-premise Exchange or Lotus
Notes

2 - There is an SLA + Enterprise Partners offer enhancements to this;
organisations are using this for 2,000 - 33,000 users so it is not
just for SME

3 - Google do not access client data - the responsibility remains with
the user - we have FS Compliance organisations using GApps and it has
passed their strict Privacy & Security requirements

4 - The change management is minimal - rolling out GApps does not take
long even when replacing multi-site Exchange or Notes implementations

5 - BPOS has major issues right now, I know because I have been using
it and have been in ongoing discussions with folks at MS; I have also
done a number of head-to-heads between BPOS and GApps so I know what
Enterprise clients think and which solution they prefer

Thanks,

J


On May 7, 12:34 pm, "Pietrasanta, Mark"

[Pietrasanta, Mark]

In terms of savings, there were a number of factors.  The gov’t is different than a lot of commercial companies in that they often hire contractors to do a lot of their work (as opposed to in-house staff and resources).

 

In this case, the previous “traditional” hosting model was a very large, physical outsourced infrastructure, *plus* a lot of services from a traditional service provider to manage it.  In the new model, shared virtualization/cloud services allow them to maintain the same capacity while significantly reduce their physical footprint (“green computing” shout out), and the reduction in complexity plus the more efficient services from the hosting provider allowed them to significantly reduce their additional managed services costs.  Also note that the saving is really big – the % is off a very large number.

 

And one more important thing, the hosting provider chosen has offered basically these same capabilities for years, long before the term “cloud computing” showed up.  This is nothing new, it’s just getting broader acceptance and gradually evolving.

 

I’m not sure this translates too well into the commercial space, and into moving from an internal data center/IT shop to a private cloud.  It seems to be more cost effective to migrate from internal traditional DC/IT towards an internal private cloud (sized for your needs, e.g. you don’t necessarily need a $400K SAN).  But I don’t have enough real data for this.

 

We ran lots of cost comparisons using the (limited cost data) from the large commercial clouds, and the costs were 4-10x or more the cost of the private cloud.  The raw hosting costs were at the lower end (4x, which alone kills it), but once you add in all the additional services work needed (since none of these are offered by the large clouds), as well as build in all the necessary items to handle the non-existent SLA and terms of use, the costs skyrocket.  Large commercial clouds still don’t make cost sense when looking at 24/7 hosting.

 

In terms of security, just like with traditional hosting models, a chunk of this is handled by the hosting provider configuration as well as the SLA (firewall rules, NIST guidelines, reporting, IDS, etc.).  The rest becomes part of our machine builds, and then we beat them up to ensure compliance.  For more secure sites, obviously this model doesn’t work and the Government will likely need to move towards an internal/closely held private cloud just to meet the much higher security requirements.

 

Let me know if you have more questions, or if I didn’t answer these adequately?

Thanks,

Mark

[Pietrasanta, Mark]

When I get back to the office, I'll pull up the list of issues our lawyers and contracts folks raised and forward the main ones. I'll also grab our cost analysis.

Can I assume you work for Google, or are a Google partner?

> -----Original Message-----
> From: cloud-c...@googlegroups.com [mailto:cloud-c...@googlegroups.com] On Behalf Of Rao Dronamraju

> Sent: Wednesday, May 06, 2009 8:35 PM
> To: cloud-c...@googlegroups.com

[Rao Dronamraju]

You did. Thank you very much.

 

Since you have done the cost benefit analysis especially in a real life large cloud implementation as mentioned below, now I can understand now why you are so emphatic about it. I will give you the benefit of the doubt.

 

Regards

[Ben Bloch]

Don't forget user experience (#5?) - particularly offline use. I love Google
apps but Office is much better experience if connection/server is slow/non
existent and generally better response/performance - for single users
anyway. Collaboration is a whole other story.

I really want to use both - depending on what I am doing (including privacy
differences) - and synch them up.

Ben

-----Original Message-----
From: cloud-c...@googlegroups.com
[mailto:cloud-c...@googlegroups.com] On Behalf Of Pietrasanta, Mark

[Nik Simpson]

There's an interesting followup to the original article here:

http://storagemojo.com/2009/05/08/cloud-sand-and-scale/

Seems that Information Week cherry picked what they wanted publish.

--
Nik Simpson

[Pietrasanta, Mark]

What do you all think of this?  I don't think I saw it discussed yet:

http://tinyurl.com/p865os

 

I've got a few issues, including mixing billing models into it, as well as the mixing of provider and consumer into the definition.

[Rao Dronamraju]

This is the first time I have seen this draft from NIST…..and never on this forum…

 

Reuven says Cloud Computing is right front and center of Federal Govt., and after almost a year or two of Cloud Computing talk in the industry NIST has come up with a “Draft” specification?....

 

When will they have the final version?....after another year or two?….does it say something about the priority attached to CC in Fed. Govt?....

 

They got their heads buried in the sand not in clouds!.

 

I think what is really needed is NOT DEFINITIONS any more. BUSINESS CASE to prove the viability of CC whether in private or public (govt.) environments.

So NIST or which ever govt. organization is responsible needs to publish a solid business case for it in the govt. and same thing should happen in the commercial environment.

Gartner, IDC and others might want to publish some thing that justifies their $XX or $XXX billion projections. Where did they get it from?...

 

One interesting thing to note from the draft definition is they mentioned advertisement based revenue model. This is the first time I have seen some one mention ad based revenue for CC.

How many corporate employees would be accessing the cloud applications after SMEs move to the clouds?...multi-millions to support the advertising based revenue model?....

or are they talking about general consumer who use applications like MS office if and when they are SaaSed in the clouds?....

 

 

---

From: cloud-c...@googlegroups.com [mailto:cloud-c...@googlegroups.com] On Behalf Of Pietrasanta, Mark

Sent: Friday, May 08, 2009 3:47 PM
To: cloud-c...@googlegroups.com

[Bernard Golden]

Think the interesting/exciting thing about the definition is not its
accuracy or thoroughness, but the fact that the Federal government is
moving very quickly and that many agencies are involved. Of course it
remains to be seen what action comes down the road, but, from
experience, the Federal government is typically way, way slow on
things.

[Bernard Golden]

Mark:

Can you clarify a few questions I have from what you've written?

1. You mention that SLA was important and that adding an SLA to a
public cloud would drive up the price. Amazon offers an SLA now, so is
the issue the level (not high enough for client needs), or lack of
ability to interact with Amazon in subsequent discussions, or were
there other factors beyond the pure service level benchmark number
(reports on issues, root cause analysis, whatever). In other words,
were there additional services the GSA or yoru firm want in the SLA
that would add costs to the base Amazon offering? Would like to
understand what the client sought in its SLA.

2. You mention that Terremark's offering isn't really anything new,
but I've seen that they tout their cloud offering and specifically its
agility. A quote I heard attributed to them was something like "It
used to take us 2 days to provision a firewall, now we can do it in
two hours." Is there no cloud functionality (quick system initiation,
scalability) offered that is visible to the end user or is the
functionality purely within Terremark's operations and they continue
to offer traditional service offerings to the client?

3. This is more for your opinion. You mention that you believe an
internal DC to cloud capability would be most cost-effective, but have
no evidence once way or the other for that. In the story, the GSA exec
quoted said something like "our folks didn't really like it because
they didn't want to give up their servers." Is it possible, in your
opinion, that despite the fact that the internal cloud might
potentially be most cost-effective, cultural issues will intrude in
those efforts and result in placing private clouds with service
providers like Terremark? I ask because I've seen many initiatives
fail not due to technical insufficiency but due to cultural issues
(e.g., this is different than what I usually do, so I have to learn to
do something new, so I'll just passively sabotage the initiative so I
can go back to the ways things were).

4. You've mentioned in previous posts that, in your experience (east
coast), most IT shops are efficiently managed and cost-effective. In
this example, the GSA seems like they're going to save a ton of money
because the previous infrastructure was inefficient and managed in a
very costly way. Is the GSA an outlier in terms of inefficiency? I
think that the benefit of cloud computing is not that it will offer
incremental efficiencies in the way things are done now (because most
IT shops feel they are being very efficient in their current
operations), but that it is a completely different way of providing IT
services, which will dramatically lower IT costs -- and unless it does
there's not much point in pursuing it. A different way of saying it is
that the only way to gain agility and scalability -- at a cost that is
tolerable -- is to move to an automated operations model that is hands-
off (which seems to be what Terremark has done).

Thanks.

On May 8, 5:06 am, "Pietrasanta, Mark" <Mark.Pietrasa...@aquilent.com>
wrote:

> In terms of savings, there were a number of factors.  The gov't is different than a lot of commercial companies in that they often hire contractors to do a lot of their work (as opposed to in-house staff and resources).
>

> In this case, the previous "traditional" hosting model was a very large, physical outsourced infrastructure, *plus* a lot of services from a traditional service provider to manage it.  In the new model, shared virtualization/cloud services allow them to maintain the same capacity while significantly reduce their physical footprint ("green computing" shout out), and the reduction in complexity plus the more efficient services from the hosting provider allowed them to significantly reduce their additional managed services costs.  Also note that the saving is really big - the % is off a very large number.

>
> And one more important thing, the hosting provider chosen has offered basically these same capabilities for years, long before the term "cloud computing" showed up.  This is nothing new, it's just getting broader acceptance and gradually evolving.
>
> I'm not sure this translates too well into the commercial space, and into moving from an internal data center/IT shop to a private cloud.  It seems to be more cost effective to migrate from internal traditional DC/IT towards an internal private cloud (sized for your needs, e.g. you don't necessarily need a $400K SAN).  But I don't have enough real data for this.
>
> We ran lots of cost comparisons using the (limited cost data) from the large commercial clouds, and the costs were 4-10x or more the cost of the private cloud.  The raw hosting costs were at the lower end (4x, which alone kills it), but once you add in all the additional services work needed (since none of these are offered by the large clouds), as well as build in all the necessary items to handle the non-existent SLA and terms of use, the costs skyrocket.  Large commercial clouds still don't make cost sense when looking at 24/7 hosting.
>
> In terms of security, just like with traditional hosting models, a chunk of this is handled by the hosting provider configuration as well as the SLA (firewall rules, NIST guidelines, reporting, IDS, etc.).  The rest becomes part of our machine builds, and then we beat them up to ensure compliance.  For more secure sites, obviously this model doesn't work and the Government will likely need to move towards an internal/closely held private cloud just to meet the much higher security requirements.
>
> Let me know if you have more questions, or if I didn't answer these adequately?
> Thanks,
> Mark
>
> From: cloud-c...@googlegroups.com [mailto:cloud-c...@googlegroups.com] On Behalf Of Rao Dronamraju
> Sent: Thursday, May 07, 2009 11:46 PM
> To: cloud-c...@googlegroups.com
> Subject: [ Cloud Computing ] Re: Are Chargeback Systems needed beforer Wider Cloud Adoption?
>
> "To help frame this up better - the Gov't moved from an external hosting provider and traditional hosting model to another hosting provider and a virtualized/*private* cloud model.  This did indeed result in substantial cost savings, and actually an *increase* in total capacity, as well as all the reliability/flexibility/etc. of moving to a private-cloud model."
>
> Just curious, how did they achieve 90% savings by moving from a traditional hosting model to another hosting provider and a virtualized /private cloud model?...
> What is in the private cloud model that contributed to such HUGE savings?....just virtualization?....
>
> "1) Large commercial clouds are completely cost prohibitive, and would have resulted in a cost *increase* to the Gov't."
>

> I am in the camp that says we do not know yet whether large commercial clouds are viable are not?....No body has enough data to prove or disprove this at this time...
>
> I think all of us in this group are debating an issues for which there just isn't enough information either way.....
>
> If you say large commercial clouds are not viable based security, compliance, governance etc....I can understand....but I am not sure about it based on cost factors....
>
> I am also interested in learning their security, compliance, governance, visibility...etc...etc issues....they have been addresses primarily through SLAs?....any specific NEW technologies deployed?....I know this could be confidential...

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

>
> On May 7, 9:52 am, "Pietrasanta, Mark" <Mark.Pietrasa...@aquilent.com>
> wrote:
> > Do you have any basis for that opinion?
>
> > In my *experience*, what you say is not consistent with actual activities.  IT shops are constantly under pressure to reduce costs, and are flooded with examples of outsourced alternatives that seem to be less expensive.
>
> > IT shops are always performing analysis of what makes the most sense
>

> ...
>
> read more »

[Johan Louwers]

Dear all,
I have been looking into something and was unable to locate it,..... at
least in the opensource community....

When we define a cloud (the term) we stated that a cloud has to be
elastic and on-demand. Now imagine a university or a big school where
students and teachers are running all kinds of projects which have the
need for computer power. Some student running a project which involves
networking so he needs 10 linux installations running, a other student
is running a project which requires him to have a server running for a
long periode and will need a lot of CPU power and memory for it. A
professor needs a lot of diskspace and some CPU power and very low
network speeds for a long periode of time for some calculations done...
you get the basic idea...

Is there already something like a portal you can see on amazone however
more simple. A user (student/professor) can request a number of systems
to be setup according to the specifications he is giving without any
(WITHOUT ANY) technical knowledge. I know that Oracle VM for example has
a web interface where you can do such things, you can setup, scale and
provision your systems however it requires quite a high security level
to get into the system and will enable you to do a lot of damage to
other running systems.

I am more looking for a controlled front end which is opensource where I
can control my own (as the user who is logged in) systems and request
more systems / storage etc ect. It would even be better if you can set
some thresholds on CPU cycles etc etc and make a department pay for the
use per (cycle / cpu / ...) and have the ability to automaticly stop the
system if the threshold is reached.

The technical maintenance and control will not have to be done via this
interface.... it is just a student facing non-tech portal which will
enable them to create / start / run / kill the systems they control. On
which server, which storage is used, etc etc etc is not of any interest
for them this is seen as a technical matter. If it becomes a matter for
them they most likely will already be attending a computer science class
and already have befriended the local IT staff and get things done that
way ;-)

I was hoping something was available in the opensource community however
I was unable to locate it ..... did I miss it or do I have to build it
and donate it to the community? If something like this turns out not to
be existing and it has to be build.... maybe this mailing list can be a
good platform to start a discussion on how to create such a thing and
what it should be providing so it can be used at a wide variate of
environments (companies / universities / .....)

Thanks already!

Regards,
Johan Louwers.