Report Offensive Message
You have not understood "the cloud"
The cloud is NOT the Internet, even if it is connected to the Internet, it will continue to work even if the connection to it is lost.
the cloud is every network or device with which you can have a user experience. All of them having computing capabilities and automartic synchronization of their data WHEN they are interconnected.
Within this cloud, you can perfectly have offline files. You have replication services, direct online or with delayed synchronization. You don't depend on a single Internet connection, you don't even have any fixed infrastructure. Any connection medium that can be available sometime can be used to transfer and consolidate the data.
The internet will then just be used as a possible transfer medium, and a possible offsite backup solution. It may be used for temporarily externalizing some service and there are solutions for this.
Just consider your phone, it is already going through the internet and most of the time you will not notice it because it works just like the trandtional analog POTS line (often it even works better in terms of audio quality), and it is going better and better (network congestions on the Internet are much less a problem; an accident can happen that could dramatically affect some area, but anyway when this happens, it not only affects the Internet, but the phone network as well; the reparation delays are almost the same and during that time you will see little difference with the situation of those organizations that already depend on their phone and fax lines, or leased lines to work with their customers and providers, or other remote offices. Basically, the risks with communications over the internet are now almost the same as with other networks.
But outages in most out-of-office network centers are MUCH LESS frequent than in most in-office server installations where failures will have more dramatic effects (due to lack of local administrators, or standardized processes for fast recovery). In-office servers also often lack the necessary costly redundancy servers that must be administered preventively in parallel, to the usual servers (but are available in standard in most out-of-office network centers, that also have several interconnected sites to hold the traffic and application charges in case of severe emergency). Many organizations don't have any secondary location where to place their servers in case of emergency, so the backup systems are often exposed to the same risks as the primary ones: connecting the organization to a network center and working with it with standardized systems will be very economic, but it does not mean that you must completely remove your in-site infrastructure. The off-site servers that you add in a network center (including those that may be available very cheaply through Internet used only as a transport medium that can be secured) allow reduced costs in many organizations (no need to maintain two separate offices), in addition to providing a cheap way to standardize and increase the accessibility of your organization (including in the smallest remote offices where it's completely impossible to budgetize the installation and maintenance of local servers). Most often, these network centers are accessible through cost-effective internet connections, secured through a VPN, that will even be much faster for everyday use than the previously leased lines or sepcific modem connections through POTS or ISDN (even the hardware router to connect to the inernet as well as the support for connecting to the VPN will be standard and supported by the ISP, with very fast reparation in cast of failure). Most organizations won't be able to repair and reconnect those remote sites as fast as an ISP.
Clearly the goal is not to replace completely the network/servers infrastructure completely by moving all the servers to a network center managed by an ISP, but to secure an existing network and give it an on-demand performance boost when needed, and increase its connectivity, and lower the adminsitration costs and delays.
We can effectively see a 99,98% availability warrantied by service providers by contract, with most failures repaired in hours, almost all upgrade/maintenance warrantied or performed during offline hours (something that would be costly for many organizations that are not even prepared with their own work force), and these warrantied installations can also be covered by business insurances (that can cover the incidental costs of the remaining possible failures, something tht wuold be impossible in many organizations where this would first require a very costly certified quality inspection).
Of course if your company is IBM or Wallmart, you can pay for your own infrastructure, but even in this case, you'll have to manage many small remote sites for which the solution of their integration in your computing infrastructure through traditional Internet accesses plus a VPN integration solution supported by this ISP can effectively reduce the costs and offer better availability and faster reparation in case of failure.
Nobody proposes to deport all your infrastructure to the same provider through the Internet, on which you will create and host all your applications, but nothing forbids you to extend and secure a significant part of it using the standardized procedures and systems supported by a large set of service providers (notably for the LAMP servers).
Anyway you must not forget that an organization does not only depend on its capability to work on its own network, but also by its interconnection with the rest of the world from which they get their customers and providers. The internet is just another mean in addition to your phones, faxes and mailings. And even your internal work force may have no other choice than using these outside networks to communicate with you: why wouldn't they also use the Internet connection to also reach your cloud, when it is cost effective, much faster and cheaper than mobile networks, and it often offers a better availability than these mobile networks?
And even if you decide to move all your storage data offsite (because you have no seriously secured management for it locally), it does not mean that an accident that would occur at your ISP will loose all your data: by contract they are standardizing the backups, and most often they are running from multiple distinct network centers on distinct places: the backups are performed across these sites. In addition, these network centers are highly secured and accessible only by specific personels, and with a 24/24 7/7 maintenance workforce, something that would be impossible to implement in many organizations where this would require new buildings, and hiring costly people, plus supporting their replacement in case of their own failure. In a network center, these personnels are effectively working for supporting the servers of several companies, but as their work is standardized, this is not mpore complicate than managing just one organization. And you'll be able to contact thier support with the same facility that you would use to contact your own local adminsitrator (except that when your local admin is in holiday, you may have to wait sometimes for the morning or the end of the week-end, and you'll need time to replace him compeltely if he's ill, even if you have some personel or contracted service providers that can perform some assistance tasks, they often won't come to help you before a few days, the time for you to contact and agree with them, or diagnose the problems and possibly having any replacement hardware available and delivered by some emergency postal service (that you'll pay a lot too in such a situation)...
The "cloud" is then not the Internet, it's how to manage the risks on YOUR OWN network using the cheap ressources offered by service providers connecting you to the Internet from now almost everywhere and thorugh various support medium. If your installed system is to be used also by your customers or your other providers, it really makes sense to locate it in a network center instead of in your office, just because it will be cheaper and will offer system availability for them, and faster responses (this is the first image that your organization will give to them).
the cloud is every network or device with which you can have a user experience. All of them having computing capabilities and automartic synchronization of their data WHEN they are interconnected.
Within this cloud, you can perfectly have offline files. You have replication services, direct online or with delayed synchronization. You don't depend on a single Internet connection, you don't even have any fixed infrastructure. Any connection medium that can be available sometime can be used to transfer and consolidate the data.
The internet will then just be used as a possible transfer medium, and a possible offsite backup solution. It may be used for temporarily externalizing some service and there are solutions for this.
Just consider your phone, it is already going through the internet and most of the time you will not notice it because it works just like the trandtional analog POTS line (often it even works better in terms of audio quality), and it is going better and better (network congestions on the Internet are much less a problem; an accident can happen that could dramatically affect some area, but anyway when this happens, it not only affects the Internet, but the phone network as well; the reparation delays are almost the same and during that time you will see little difference with the situation of those organizations that already depend on their phone and fax lines, or leased lines to work with their customers and providers, or other remote offices. Basically, the risks with communications over the internet are now almost the same as with other networks.
But outages in most out-of-office network centers are MUCH LESS frequent than in most in-office server installations where failures will have more dramatic effects (due to lack of local administrators, or standardized processes for fast recovery). In-office servers also often lack the necessary costly redundancy servers that must be administered preventively in parallel, to the usual servers (but are available in standard in most out-of-office network centers, that also have several interconnected sites to hold the traffic and application charges in case of severe emergency). Many organizations don't have any secondary location where to place their servers in case of emergency, so the backup systems are often exposed to the same risks as the primary ones: connecting the organization to a network center and working with it with standardized systems will be very economic, but it does not mean that you must completely remove your in-site infrastructure. The off-site servers that you add in a network center (including those that may be available very cheaply through Internet used only as a transport medium that can be secured) allow reduced costs in many organizations (no need to maintain two separate offices), in addition to providing a cheap way to standardize and increase the accessibility of your organization (including in the smallest remote offices where it's completely impossible to budgetize the installation and maintenance of local servers). Most often, these network centers are accessible through cost-effective internet connections, secured through a VPN, that will even be much faster for everyday use than the previously leased lines or sepcific modem connections through POTS or ISDN (even the hardware router to connect to the inernet as well as the support for connecting to the VPN will be standard and supported by the ISP, with very fast reparation in cast of failure). Most organizations won't be able to repair and reconnect those remote sites as fast as an ISP.
Clearly the goal is not to replace completely the network/servers infrastructure completely by moving all the servers to a network center managed by an ISP, but to secure an existing network and give it an on-demand performance boost when needed, and increase its connectivity, and lower the adminsitration costs and delays.
We can effectively see a 99,98% availability warrantied by service providers by contract, with most failures repaired in hours, almost all upgrade/maintenance warrantied or performed during offline hours (something that would be costly for many organizations that are not even prepared with their own work force), and these warrantied installations can also be covered by business insurances (that can cover the incidental costs of the remaining possible failures, something tht wuold be impossible in many organizations where this would first require a very costly certified quality inspection).
Of course if your company is IBM or Wallmart, you can pay for your own infrastructure, but even in this case, you'll have to manage many small remote sites for which the solution of their integration in your computing infrastructure through traditional Internet accesses plus a VPN integration solution supported by this ISP can effectively reduce the costs and offer better availability and faster reparation in case of failure.
Nobody proposes to deport all your infrastructure to the same provider through the Internet, on which you will create and host all your applications, but nothing forbids you to extend and secure a significant part of it using the standardized procedures and systems supported by a large set of service providers (notably for the LAMP servers).
Anyway you must not forget that an organization does not only depend on its capability to work on its own network, but also by its interconnection with the rest of the world from which they get their customers and providers. The internet is just another mean in addition to your phones, faxes and mailings. And even your internal work force may have no other choice than using these outside networks to communicate with you: why wouldn't they also use the Internet connection to also reach your cloud, when it is cost effective, much faster and cheaper than mobile networks, and it often offers a better availability than these mobile networks?
And even if you decide to move all your storage data offsite (because you have no seriously secured management for it locally), it does not mean that an accident that would occur at your ISP will loose all your data: by contract they are standardizing the backups, and most often they are running from multiple distinct network centers on distinct places: the backups are performed across these sites. In addition, these network centers are highly secured and accessible only by specific personels, and with a 24/24 7/7 maintenance workforce, something that would be impossible to implement in many organizations where this would require new buildings, and hiring costly people, plus supporting their replacement in case of their own failure. In a network center, these personnels are effectively working for supporting the servers of several companies, but as their work is standardized, this is not mpore complicate than managing just one organization. And you'll be able to contact thier support with the same facility that you would use to contact your own local adminsitrator (except that when your local admin is in holiday, you may have to wait sometimes for the morning or the end of the week-end, and you'll need time to replace him compeltely if he's ill, even if you have some personel or contracted service providers that can perform some assistance tasks, they often won't come to help you before a few days, the time for you to contact and agree with them, or diagnose the problems and possibly having any replacement hardware available and delivered by some emergency postal service (that you'll pay a lot too in such a situation)...
The "cloud" is then not the Internet, it's how to manage the risks on YOUR OWN network using the cheap ressources offered by service providers connecting you to the Internet from now almost everywhere and thorugh various support medium. If your installed system is to be used also by your customers or your other providers, it really makes sense to locate it in a network center instead of in your office, just because it will be cheaper and will offer system availability for them, and faster responses (this is the first image that your organization will give to them).
Posted by PhilippeV
13th Jul 2009
