Open Source Cloud Computing: Solving privacy issues (eyeOS Project)
11 views
Skip to first unread message
Pau Garcia-Milà
Jan 29, 2009, 10:14:43 AM1/29/09
to cloud-c...@googlegroups.com
Every day I see new people discovering the "Cloud Computing". In a first view at the concept it sounds cool -data ubiquity, accessing everything from everywhere, backing up at the cloud....- but after thinking a bit on it, the privacy concern appears: We are sending our information to someone else's servers. And we depend on Internet connectivity to access all our digital life and business files.
I really think that the next big monopoly in the IT industry will be for that company who will control the data from 90% of users and companies. And that will be possible because of Cloud Computing, which should worry us. We believe in Cloud Computing, but we want it to be free, without needing to send away our digital lifes.
Some of you probably already know about the eyeOS Project. It wants to join Cloud Computing and Open Source software to solve the privacy problems around the Cloud. eyeOS is a Free Software project that companies and particulars install on their servers -so data and apps are kept under control- and they can work with them, through a web browser, from everywhere, inside and outside the company/home. It provides some base apps (to work/create MS Office and OpenOffice compatible documents, calendar, contacts, mail, RSS...) and a Toolkit to develop new apps on. It currently holds 9 communities in 9 countries and around half a milion downloads. And it has been developed since 2005 -some time before we started to hear about Cloud Computing- .
Just wanted to share with you the project info and ideas. More info is at its website, www.eyeos.org
Thanks!
I really think that the next big monopoly in the IT industry will be for that company who will control the data from 90% of users and companies. And that will be possible because of Cloud Computing, which should worry us. We believe in Cloud Computing, but we want it to be free, without needing to send away our digital lifes.
Some of you probably already know about the eyeOS Project. It wants to join Cloud Computing and Open Source software to solve the privacy problems around the Cloud. eyeOS is a Free Software project that companies and particulars install on their servers -so data and apps are kept under control- and they can work with them, through a web browser, from everywhere, inside and outside the company/home. It provides some base apps (to work/create MS Office and OpenOffice compatible documents, calendar, contacts, mail, RSS...) and a Toolkit to develop new apps on. It currently holds 9 communities in 9 countries and around half a milion downloads. And it has been developed since 2005 -some time before we started to hear about Cloud Computing- .
Just wanted to share with you the project info and ideas. More info is at its website, www.eyeos.org
Thanks!
Miha Ahronovitz
Jan 30, 2009, 2:32:36 AM1/30/09
to cloud-c...@googlegroups.com
Pau, this is beautiful , but is this a cloud? A couple of us here, on this group , support this definition of the cloud
A person attracted to the cloud after reading all the buzz, has two key expectations:
1. The ability to get a quality service any time (how the provider
will get in minutes the resources to provide this quality is not his/
her business)
2. The ability to pay only for what s/he uses when s/he needs it.
The rest is an implementation details. Users want to be totally free
away from any technical complexity other than the service itself.
In the eyeOS.org, I will own my own server, and I need more resources, do I need to go an buy another server or I can rent (and pay) from another place a server, virtual or physical ? I understand the eyeOS community does not offer access to a pool of stand-by servers or servers which are idle elsewhere.
2nd, the users have all the headache to configure the server. This is not straightforward, you offer a myriad of partners who can configure my server, i.e. my foothold in the cloud. This is not a Kibbutz, I own my server. Then you offer to write the applications for each customer who joins eyeOS.
http://eyeos.com/en/installation
http://eyeos.com/en/development
Third, users don’t pay only for what they use. Each participant pays for his server, whether he uses it or not. Probably at any given moment, there are thousands of individually owned servers idle or underutilized. Who pays for them? Every individual owner.
Pau, I think you are close to having a cloud, if you change the business model and have a company own the servers and manage them and places all applications. Then that company offers a worry free service to each member, and you will have a cloud with substantial economies of scale.
There are other ways to keep the security of data inside the cloud, other than keeping everything on my own server. If some users will be insisting in having a local data storage, the cloud company can charge an additional service and back up daily or hourly all data from that account to the owner’s server.
2 cents,
Miha
Pau Garcia-Milà
Jan 30, 2009, 9:53:05 AM1/30/09
to Cloud Computing
Hi Miha,
Thanks for taking some time to analyize it. I think we agree that the
"Cloud Computing" definition is really undefined. And I'm really
conscient that with eyeOS we just make it a bit more ambiguous. In
fact we do provide servers commercially and there is a free demo
server of eyeos which has about 400.000 users at eyeos.info, but the
key point is the ability to host it -just if the company needs/wants
it-.
For example, for a small company/group I would tell them to pick an
eyeOS VPS server where we will give them the virtual server hosting,
so they won't need to build a cloud server. But probably for a larger
company, administration or university sending away their data won't be
an option. And there is where we want to focus: To provide a solution
that can (if the user wants it) be installed in an existing server /
servers farm.
Our objective is to create an Open Source Solution where cloud
applications can be built on. And once created, can be hosted
anywhere.
Regards from Barcelona,
Pau Garcia-Milà
Thanks for taking some time to analyize it. I think we agree that the
"Cloud Computing" definition is really undefined. And I'm really
conscient that with eyeOS we just make it a bit more ambiguous. In
fact we do provide servers commercially and there is a free demo
server of eyeos which has about 400.000 users at eyeos.info, but the
key point is the ability to host it -just if the company needs/wants
it-.
For example, for a small company/group I would tell them to pick an
eyeOS VPS server where we will give them the virtual server hosting,
so they won't need to build a cloud server. But probably for a larger
company, administration or university sending away their data won't be
an option. And there is where we want to focus: To provide a solution
that can (if the user wants it) be installed in an existing server /
servers farm.
Our objective is to create an Open Source Solution where cloud
applications can be built on. And once created, can be hosted
anywhere.
Regards from Barcelona,
Pau Garcia-Milà
satish rege
Jan 30, 2009, 1:35:39 PM1/30/09
to cloud-c...@googlegroups.com
Paul,
Thanks for the post. I have not spent time looking at eyeOS. I will do that. But I was curious if it solves data "privacy concerns" that you mentioned as an important consideration in your first post.
-satish
Thanks for the post. I have not spent time looking at eyeOS. I will do that. But I was curious if it solves data "privacy concerns" that you mentioned as an important consideration in your first post.
-satish
Miha Ahronovitz
Jan 30, 2009, 2:10:36 PM1/30/09
to cloud-c...@googlegroups.com
> I think we agree that the "Cloud Computing" definition is really
undefined.
If it is undefined, how can we qualify what is cloud computer and what not?
undefined.
How can we qualify eyesOS in its present incarnation as a cloud?
One of the positive outcome from joining this group was -for me- to
understand what cloud computing is from a business point of view.
What definition shall we use if we want to make money from such a cloud?
So the result is the definition below: ability to run an application always
and paying only for I use and it is easy to use.
Let's compare eyeOS with say Yahoo or Google.
1. I have a failure: I don't know if it is on my server or in the eyeOS
"cloud". In google I don't care
2. I have too many users and I need temporarily more resources. I need to
buy a server,
or call by phone to get a virtual server. In google or Amazon all is
automatic.
As long as I have to do additional administrative work outside the "cloud",
this is not a cloud.
Assume I get funding to start a new cloud venture offering services: I want
to attract as many users
who join as simple as making a google query. Then I want them to have
immediate response,
no matter how many paying customers I have during the peak. All they need is
to access the cloud,
no more headaches, and to pay a bill that has no surprise charges
Stacey Schneider from this group published the IDC definitions for a cloud,
See them after the sig line
Note the cloud services definitions. The definitions are not yet practical,
IMHO,
but they point out to the simple definition a few of us agree as useful,
if I want to make a cloud a financial success for their funders and
investors.
The eyesOS is beautiful project, and advanced proto-cloud, one step away
becoming one, .
I predict it's commercial success, if it switches the focus of making money
from
professional services.
Cheers,
miha
> IDC makes this distinction:
>
> * *Cloud services* are the consumer and business products,
> services, and solutions that are delivered and consumed in real time
> over the Internet. (IDC identifies eight key attributes that more
> clearly define the new generation of commercial cloud services.)
>
> * *Cloud computing* is an emerging IT development, deployment, and
> delivery model, enabling real-time delivery of products, services, and
> solutions over the Internet.
>
> The former is a business model, the latter is an IT technology.
>
> Cloud Services..
>
> o lower costs,
> o simplify and accelerate access
> o enable fine-tuned provisioning,
> o greatly increase the number and variety of available services, and
> o improve the potential to integrate these services.
>
> Cloud computing consists of a growing list of technologies and IT
> offerings that enable cloud services, including
>
> o infrastructure systems (servers, storage, networks),
> o application software,
> o system and application management software,
> o IP networks, and
> o pricing agreements.
>
> (Interesting the pricing agreements are listed alongside infrastructure)
Pau Garcia-Milà
Jan 30, 2009, 4:28:48 PM1/30/09
to Cloud Computing
Hi Miha,
I would say is undefined because not 2 people here would define it in
the same way, so I'm not the one to tell which one is correct and
which one not. I agree on your definition of using X apps and just
paying for the resources I use, but I think that's not a valid option
for some places (e.g. a public administration) if its outsided. So
isn't cloud computing for them? I think it's for them, but they need
their "internal cloud", even if that means having to pay for the
resources. Maybe I'm trying to expand or modify the cloud's definition
here.
More on that, I think every closed system needs a free software one.
What Linux is to Windows. Hope eyeOS (or another project, but free,
open source software) will grow enough to be the free one to microsoft
& google ones. I really think the world needs that, since an important
sector's opinion is that the only way to keep 100% privacy on our
files is controling them in our servers. And we want to make a cloud
computing "operating system" for them too. If that's not such a
problem for some others, they can just pick a hosted eyeOS server --we
offer them! But always maintaing free the source code so for those who
want it can build their cloud.
However, will take your advice seriously for sure.
Satish,
Please let me know your opinion after checking it ;-)
Thx,
Pau
I would say is undefined because not 2 people here would define it in
the same way, so I'm not the one to tell which one is correct and
which one not. I agree on your definition of using X apps and just
paying for the resources I use, but I think that's not a valid option
for some places (e.g. a public administration) if its outsided. So
isn't cloud computing for them? I think it's for them, but they need
their "internal cloud", even if that means having to pay for the
resources. Maybe I'm trying to expand or modify the cloud's definition
here.
More on that, I think every closed system needs a free software one.
What Linux is to Windows. Hope eyeOS (or another project, but free,
open source software) will grow enough to be the free one to microsoft
& google ones. I really think the world needs that, since an important
sector's opinion is that the only way to keep 100% privacy on our
files is controling them in our servers. And we want to make a cloud
computing "operating system" for them too. If that's not such a
problem for some others, they can just pick a hosted eyeOS server --we
offer them! But always maintaing free the source code so for those who
want it can build their cloud.
However, will take your advice seriously for sure.
Satish,
Please let me know your opinion after checking it ;-)
Thx,
Pau
Pau Garcia-Milà
Jan 30, 2009, 4:54:35 PM1/30/09
to Cloud Computing
Just as an extra to the discussion, an article featured in Linux
Journal (January 2009 issue) about eyeOS + Clouds:
http://www.linuxjournal.com/content/eyeos-clouds-crowd
I think that article clearly explains my point of view
Thanks
Journal (January 2009 issue) about eyeOS + Clouds:
http://www.linuxjournal.com/content/eyeos-clouds-crowd
I think that article clearly explains my point of view
Thanks
satish rege
Jan 30, 2009, 7:01:56 PM1/30/09
to cloud-c...@googlegroups.com
Paul,
Seems like your solution is to keep the "cloud" within the intranet to solve the privacy issue. (I have still not read eyeOS, but gather my inference from your comments).
What ever definition for Cloud Computing is finally agreed upon by the community, one of the main goals of Cloud Computing is to substantially reduce (minimize) cost, while providing higher level of reliability and availability. An intranet cloud can reduce cost only up to a point, because its resources are generally used by less number of users, or in other words it cannot accrue the benefits of large population of users sharing resources. Note also that large population of users and their needs has a property of smoothing out resource utilization, which helps an internet cloud.
But as you say, with the present technology, if privacy issue is important or if you are paranoid, then you have to keep your data in house paying more for running your application.
I will share an interesting statistics that was accepted a few years ago and is still true. For $1 that is spent on buying storage, $7 - $10 are spent on maintaining data residing on that storage per year. I do not know the costs of maintaining, enhancing CPU, operating systems and applications.
My final question is - can we come up with a solution to solve the privacy issue over the internet?
By the way, I do like the idea of open source eyeOS though. Don't miss-understand me on that.
Wish you all the luck.
-satish
Seems like your solution is to keep the "cloud" within the intranet to solve the privacy issue. (I have still not read eyeOS, but gather my inference from your comments).
What ever definition for Cloud Computing is finally agreed upon by the community, one of the main goals of Cloud Computing is to substantially reduce (minimize) cost, while providing higher level of reliability and availability. An intranet cloud can reduce cost only up to a point, because its resources are generally used by less number of users, or in other words it cannot accrue the benefits of large population of users sharing resources. Note also that large population of users and their needs has a property of smoothing out resource utilization, which helps an internet cloud.
But as you say, with the present technology, if privacy issue is important or if you are paranoid, then you have to keep your data in house paying more for running your application.
I will share an interesting statistics that was accepted a few years ago and is still true. For $1 that is spent on buying storage, $7 - $10 are spent on maintaining data residing on that storage per year. I do not know the costs of maintaining, enhancing CPU, operating systems and applications.
My final question is - can we come up with a solution to solve the privacy issue over the internet?
By the way, I do like the idea of open source eyeOS though. Don't miss-understand me on that.
Wish you all the luck.
-satish
Pau Garcia-Milà
Jan 31, 2009, 5:48:01 AM1/31/09
to Cloud Computing
Thx for your comments, Miha and Satish.
I really understand the cloud computing definition from a user's point
of view, but I think anyone should be able to create an own cloud
system. And that's the point from eyeOS.
As a note, usually our customers (governments in Spain, some companies
in France and USA) usually switch to eyeOS to reduce costs from other
technologies like traditional virtualization systems. They reduce
license costs, servers costs -you can run about 10x simultaneous users
in an eyeos server than in a virtualization server, because of using
just a web server to run-, and terminal costs. They do not want to use
our hosting options (even we offer them!) because its a need for them
to keep the data "inside the building", but they want to centralize
data and apps in a single point, offer their employees and maybe
customers a cloud solution. In that case I really think eyeOS is cloud
computing, it's still centralized but from inside the company.
Again (sorry for repeating the same) user's can also use eyeOS from
eyeos.info server (which has about 400.000 accounts), and choose an
eyeOS virtual private servers with us, but this is just another cloud
option.
Please do not leave the discussion, I'm really interested in fitting
the cloud definition and not making it more ambiguous.
> ...
>
> read more »
I really understand the cloud computing definition from a user's point
of view, but I think anyone should be able to create an own cloud
system. And that's the point from eyeOS.
As a note, usually our customers (governments in Spain, some companies
in France and USA) usually switch to eyeOS to reduce costs from other
technologies like traditional virtualization systems. They reduce
license costs, servers costs -you can run about 10x simultaneous users
in an eyeos server than in a virtualization server, because of using
just a web server to run-, and terminal costs. They do not want to use
our hosting options (even we offer them!) because its a need for them
to keep the data "inside the building", but they want to centralize
data and apps in a single point, offer their employees and maybe
customers a cloud solution. In that case I really think eyeOS is cloud
computing, it's still centralized but from inside the company.
Again (sorry for repeating the same) user's can also use eyeOS from
eyeos.info server (which has about 400.000 accounts), and choose an
eyeOS virtual private servers with us, but this is just another cloud
option.
Please do not leave the discussion, I'm really interested in fitting
the cloud definition and not making it more ambiguous.
>
> read more »
Jordi Torres (Universitat Politècnica de Catalunya, UPC)
Jan 31, 2009, 1:56:53 PM1/31/09
to cloud-c...@googlegroups.com
Hi, I'm preparing a master course for this quarter that will include cloud
computing. Very very interesting discussion about cloud definition. Thank
you for that.
In my opinion we've defined/redefined cloud computing to include everything
that we already do before it appeared (Cloud computing search trend)
http://farm4.static.flickr.com/3093/2627760300_c168cbde13_o.jpg
Cloud Computing became a buzzword even before it was defined!
Could the OyesOS's view fit with the Sun's classification?
http://de.sun.com/emrkt/innercircle/newsletter/1108/feature-itm.html?cid=e73
41f
The Private clouds in the "Types of deployment" dimension?
Regards,
Jordi.
--
http://www.JordiTorres.org
-----Mensaje original-----
De: cloud-c...@googlegroups.com
[mailto:cloud-c...@googlegroups.com] En nombre de Pau Garcia-Milà
Enviado el: sábado, 31 de enero de 2009 11:48
Para: Cloud Computing
Asunto: [ Cloud Computing ] Re: Open Source Cloud Computing: Solving privacy
computing. Very very interesting discussion about cloud definition. Thank
you for that.
In my opinion we've defined/redefined cloud computing to include everything
that we already do before it appeared (Cloud computing search trend)
http://farm4.static.flickr.com/3093/2627760300_c168cbde13_o.jpg
Cloud Computing became a buzzword even before it was defined!
Could the OyesOS's view fit with the Sun's classification?
http://de.sun.com/emrkt/innercircle/newsletter/1108/feature-itm.html?cid=e73
41f
The Private clouds in the "Types of deployment" dimension?
Regards,
Jordi.
--
http://www.JordiTorres.org
-----Mensaje original-----
De: cloud-c...@googlegroups.com
[mailto:cloud-c...@googlegroups.com] En nombre de Pau Garcia-Milà
Enviado el: sábado, 31 de enero de 2009 11:48
Para: Cloud Computing
Asunto: [ Cloud Computing ] Re: Open Source Cloud Computing: Solving privacy
Miha Ahronovitz
Jan 31, 2009, 2:26:12 PM1/31/09
to cloud-c...@googlegroups.com
Pau, as first step you may offer additional virtual nodes
on-demand to users who exceed the capacity of their local
servers.
We will have this feature in Sun Grid Engine soon.
See here a demo by our partners Bio-Team and Univa
http://gridengine.info/2008/07/14/univa-ud-unicluster-express-and-amazon-ec2
This is what we called a hybrid cloud in many posts.
miha
on-demand to users who exceed the capacity of their local
servers.
We will have this feature in Sun Grid Engine soon.
See here a demo by our partners Bio-Team and Univa
http://gridengine.info/2008/07/14/univa-ud-unicluster-express-and-amazon-ec2
This is what we called a hybrid cloud in many posts.
miha
Pau Garcia-Milà
Feb 2, 2009, 5:00:09 AM2/2/09
to Cloud Computing
I think this debate evolved from presenting eyeOS as a Cloud solution
to the Cloud Computing definition itself (which is quite interesting).
From my point of view, Cloud Computing it's not directly related to
the size of the network that provides the services. With this I mean
that a scalable server (e.g. with a ZFS filesystem and 2 hard drives)
of 500$ from Dell can be a Cloud provider for a network of 50 people,
and can be scaled transparently to the end users in multiple ways. And
from that point, eyeOS can be a good Cloud solution for those non-
Google and not-Amazon companies.
What I tried to tell with all this thread was that the vision behind
eyeOS is to allow anyone to develop cloud solutions with a solid
backend (eyeOS+eyeOS Toolkit) without needing an investment for
servers or moving the whole data base to services like Amazon S3. So
they are not cloud storage users, but "full" cloud computing
providers.
Imaging a world of Cloud Computing with thousands of solutions where
99% of data is controlled by less than 10 companies (like Amazon with
S3, google with Google Apps, probably Microsoft, Yahoo...) should
worry us. Maybe they have lots of resources, but we all know that a
market controlled by few giant companies isn't a good place to
investigate and create. And this is why we created eyeOS.
On 31 Gen, 20:26, Miha Ahronovitz <mij...@sbcglobal.net> wrote:
> Pau, as first step you may offer additional virtual nodes
> on-demand to users who exceed the capacity of their local
> servers.
>
> We will have this feature in Sun Grid Engine soon.
> See here a demo by our partners Bio-Team and Univa
>
> http://gridengine.info/2008/07/14/univa-ud-unicluster-express-and-ama...
> ...
>
> llegiu-ne més »
to the Cloud Computing definition itself (which is quite interesting).
From my point of view, Cloud Computing it's not directly related to
the size of the network that provides the services. With this I mean
that a scalable server (e.g. with a ZFS filesystem and 2 hard drives)
of 500$ from Dell can be a Cloud provider for a network of 50 people,
and can be scaled transparently to the end users in multiple ways. And
from that point, eyeOS can be a good Cloud solution for those non-
Google and not-Amazon companies.
What I tried to tell with all this thread was that the vision behind
eyeOS is to allow anyone to develop cloud solutions with a solid
backend (eyeOS+eyeOS Toolkit) without needing an investment for
servers or moving the whole data base to services like Amazon S3. So
they are not cloud storage users, but "full" cloud computing
providers.
Imaging a world of Cloud Computing with thousands of solutions where
99% of data is controlled by less than 10 companies (like Amazon with
S3, google with Google Apps, probably Microsoft, Yahoo...) should
worry us. Maybe they have lots of resources, but we all know that a
market controlled by few giant companies isn't a good place to
investigate and create. And this is why we created eyeOS.
On 31 Gen, 20:26, Miha Ahronovitz <mij...@sbcglobal.net> wrote:
> Pau, as first step you may offer additional virtual nodes
> on-demand to users who exceed the capacity of their local
> servers.
>
> We will have this feature in Sun Grid Engine soon.
> See here a demo by our partners Bio-Team and Univa
>
>
> llegiu-ne més »
Reply all
Reply to author
Forward
0 new messages