Cloud computing appeared in the mid of 2007 according to Google trends, which obviously is not the pure truth, but we can pretty much agree that during first half of 2007 the term started to be known.
Cloud computing is the name of very old technology that allows getting services from a remote c0mputing centers. In the 60's the computers were attached with punch cards, cards readers and large rooms full of computers, while now it is virtual machine based operation.
Two ideas are new after all and should pretty much be taken in account:
- X-As-A-Service Users need services - So companies should offer the users services and not software. In most of fields we use services while in computers (due to some unclear reason), companies offer software. You do not ask a GSM software for your cellular and you should neither ask for word processor software. You might need a word processor - that is fine, but users do not need the software, but its functionality. X can be one or more of the following: software, platform, infrastructure, operation, etc. that is given as a service.
- Decoupling between business and IT. This is (IMHO) the biggest change since the appearance of the PC. Companies should no longer look at the IT as wired department that do "some computer work that only they understand". The IT will have to look around and see how it can contribute to the business of the company.
Approx 2 month ago, I asked on LinkedIn What is the "right" place of the IT in an organization ? I wanted to get an estimation of how people see the IT now in order to compare it in the future. All the replies talk about the IT as very distanced from the business. This will change in the era of "cloud computing". If a company uses SaaS or PaaS, the IT will have to invent new roles and get into other business. More of the same is not an option. Given that most of the software can already be used on the net as Office (e.g. google docs), CRM (sales.com) and many others, obsoletes the routine installation and upgrade process that some IT departments spend most of the time on updates and alike.
On one hand, the IT workers are skillful in implementing the same routine again and again and they will have to adapt to the new era. This time each IT worker will be required to have some understanging of the business they work for. If an IT worker in a bank will not be a mini-banker, his contribution to the organization will be less efficient.
Not everything can be moved to cloud. I some cases it is due to technical problems (as the organizations can not allow all their information to be external) and in some cases the IT workers will make huge efforts to prevent it. In either cases, this cloud buzz will force them to learn new practicies. In the future, it might happen that someone will suggest to move the main compute power from the serivce centers to the users. It will not be a desktop but rahter some lite weight device (maybe the cellular successor) and the computers marker will continue to make money out of those transitions.
Sunday, December 21, 2008
Tuesday, July 22, 2008
What missing in the gird world ? - Applications!
If one wonders why the grid is not widely used, the answer is trivial - A p p l i c a t i o n s!
Obvious - not ?
Well not so obvious. There are many other issues that cause the application to be not present.
First is the parallel thinking, most of us think serial, or at least not really parallel. It causes both, the application development and the user understanding more complicated. So it should be covered with a lot of abstraction layers that will give good feelig for the user.
Developers, QA, trainers, etc. need to change the way they think. It is not easy to make this major change mainly since there is no methodologies and best practices. The lack of tools can be seen as great opportunity. There is much knowledge in this area, but it is concentrated within the universities and research institutes. Those organizaitons do not understand the industry needs nor the business models.
What about the need?
It is clear that more compute power can be used and contribute to any company, on the other side it is not clear what is the marginal contrubution. The right question qould be not what we gain from doing it but rather what we loose if we do not do it. That is the exact problem of the grid applications. There are many killer apps, but none of them was found yet....
Developing the business model for grid application will probably make a difference. In the past, search engines did not had a common incoming model. Then came Google and the rest is history. SaaS, cloud, etc. are nice ideas, but they are not the business model that you will crash on.
We wait very much for the Google of the Grid world.
Obvious - not ?
Well not so obvious. There are many other issues that cause the application to be not present.
First is the parallel thinking, most of us think serial, or at least not really parallel. It causes both, the application development and the user understanding more complicated. So it should be covered with a lot of abstraction layers that will give good feelig for the user.
Developers, QA, trainers, etc. need to change the way they think. It is not easy to make this major change mainly since there is no methodologies and best practices. The lack of tools can be seen as great opportunity. There is much knowledge in this area, but it is concentrated within the universities and research institutes. Those organizaitons do not understand the industry needs nor the business models.
What about the need?
It is clear that more compute power can be used and contribute to any company, on the other side it is not clear what is the marginal contrubution. The right question qould be not what we gain from doing it but rather what we loose if we do not do it. That is the exact problem of the grid applications. There are many killer apps, but none of them was found yet....
Developing the business model for grid application will probably make a difference. In the past, search engines did not had a common incoming model. Then came Google and the rest is history. SaaS, cloud, etc. are nice ideas, but they are not the business model that you will crash on.
We wait very much for the Google of the Grid world.
Friday, July 4, 2008
Middle-ware - the glue between companies
Now it is clear, the computer resources are relatively cheap when using it as part of the Terra-computing. If Amazon can offer storage and CPU time lower then the cost of disks and retail computer price while it includes installation and maintenance (for storage it probalby includes backup too), the "Cloud computing" is here.
This architecture changes some of the concepts as can be see here. This time, I want to focus on the meaning of the "middleware" and its role in this environment. According to Wikipedia, it can be seen as "services that allow multiple processes running on one or more machines to interact across a network". So in this case the middleware will connect between resources that are located in different autonomous systems. Wait a moment....that is one of the grid definitions. So do we expect the middleware to connect storage from S3 at Amazon with Google-App-Engine ?
It is simple - YES !
The middleware would need to allow creating collaboration over outsourced resources! In the future we might see the cloud computer providers as telephone companies. Huge companies that offer services cheap. But still the middleware will create different clouds for the programmers and will allow focusing on the business intelligence rather than on the maintenance.
Will it be free ? - I hope not always!
Since the commercial companies (and other users) might need a legal contract (that states privacy), this service should not be always free. On the other hand, and as always, new models will come up. Today the most ubiquitous model includes commercials. As in Hotmail. Gmail. etc. - you do not pay money, but you see our commercials. IMHO more models would be seen.
This architecture changes some of the concepts as can be see here. This time, I want to focus on the meaning of the "middleware" and its role in this environment. According to Wikipedia, it can be seen as "services that allow multiple processes running on one or more machines to interact across a network". So in this case the middleware will connect between resources that are located in different autonomous systems. Wait a moment....that is one of the grid definitions. So do we expect the middleware to connect storage from S3 at Amazon with Google-App-Engine ?
It is simple - YES !
The middleware would need to allow creating collaboration over outsourced resources! In the future we might see the cloud computer providers as telephone companies. Huge companies that offer services cheap. But still the middleware will create different clouds for the programmers and will allow focusing on the business intelligence rather than on the maintenance.
Will it be free ? - I hope not always!
Since the commercial companies (and other users) might need a legal contract (that states privacy), this service should not be always free. On the other hand, and as always, new models will come up. Today the most ubiquitous model includes commercials. As in Hotmail. Gmail. etc. - you do not pay money, but you see our commercials. IMHO more models would be seen.
Subscribe to:
Posts (Atom)