The first version of this article was written back in 2005 as a contribution to a strategic planning exercise. I was arguing for putting those things like marketing and so on, that we ultimately wanted to be seen, in the Cloud as this would save money. But I stressed that the Cloud is not secure so it was imperative that we retained our 'physical corporate information infrastructure'.
I predicted that in future: 'Corporations will need to look to ways in which they can split their computing requirements between the internal systems and the Cloud'; and 'We can expect the development of new software systems to facilitate (e-mail attached) document passing between secure corporate environments and the Cloud'; implicitly, because e-mail is not secure.
The article was slightly updated in 2010, when I opened this website. I added some now outdated and since removed comments on that experience.
I've left the main content here, essentially unchanged since 2005, because it's still relevant; and it's always good to see one's predictions fulfilled.
This Website is hosted in the Cloud. This has a lot of advantages over conventional web hosting.
But is the Cloud the solution of the future for all users? In particular is it a business solution for larger business and government?
In the last is 30 years computing has come a long way. Prior to 1980 most business computing was mainframe based and the most sophisticated desktop machine in most offices was an IBM Selectric typewriter.
Very few corporations supported networked desktop workstations. By the end of the next decade corporate networking, using workstations, was ubiquitous but external connections were few. The Internet was in its infancy and much of its content dubious.
But in the past 10 years the Internet has become the major marketing arm of most businesses. There has been a rapid expansion in the number of websites and the development of internal intranets; e-mail has become the predominant means of corporate communication.
Most informed commentators believe that these trends will continue.
Since the advent of the Internet it has become traditional to symbolise it in network diagrams as a Cloud and this has become a metaphor for computing services that can be delivered to a user by means of connection to the World Wide Web.
Cloud computing anticipates that users will purchase services rather than the enabling infrastructure and capital items. For example, a user may purchase a mobile device (eg phone) plan which will provide them with services that are drawn from the Cloud. These might be communications, information services, entertainment, or computing applications such as word processing and accounting. Consumers may not have, or own, their own storage capacity but keep their photographs, documents and accounts on the server (eg on Face Book).
There are obvious advantages. A user can connect to their documents and information anywhere in the world from any suitable device. They can move from device to device and can upgrade easily. They can form communities of interest and easily share content (eg media) and information with others. The Cloud facilitates instant messaging and ‘chat’ and provides a medium for publishing a vast amount of information including ‘wikis’, global maps, instantaneous weather information, stock market trades and other news media. Informal news services and ‘blogs’ have been made possible.
Correspondingly, many new businesses have grown up on the back of these opportunities and these have in turn resulted in considerable economic growth and expansion. A whole new generation of consumers is being developed for whom a continuous connection and instantaneous chat are essential elements of modern life. Expansion of the Cloud paradigm appears to be a trend that will continue for many years to come.
The term virtualisation encompasses a number of technologies that facilitate Cloud computing. In its broadest sense virtualisation makes one device or software application emulate another. A group of servers might appear to be a single server; one operating system might appear to be another (an Apple Mac or Linux box may appear to be running Windows or vice versa); a WorkStation maybe may appear to be word processing locally when in fact it is happening on a remote server.
Of course none of this happens magically. In each case there is a software application which provides an interface between the real underlying infrastructure and the virtual. And in each case there is a cost. This might include the need for additional processing or memory capacity and very likely, the addition of licence fees to a developer.
Corporations can be viewed as a collection of individuals and it is possible to imagine the Corporation of the future operating without any physical corporate information infrastructure of its own.
In this hypothetical Corporation every employee would simply connect to the Cloud by whatever means were available. Even the interface devices would be provided free, the cost being wrapped into the service fees.
All corporate information would be stored on host or servers somewhere in the World. As data may be distributed over many data storages, in many locations anywhere in the world, just where the user would not know. But the Corporation would have the assurance of their service provider that their information was safe due to multiple redundancy and other advanced features.
The only information infrastructure the Corporation would need to retain would be the internal knowledge management expertise that was essential to the market differentiation of their products or services and to the operation of the business.
While this is a vision held by many vendors in the computing services world, it has a number of serious shortcomings at the present time.
Principal amongst these are security, trust, performance and cost.
No computer system is completely secure and there is considerable comfort in having direct control and oversight of the physical storage of data. Security is further compromised every time data is moved from one storage medium or location to another.
WWW might stand for World Wild Web. In addition to human hackers, a vast array of software crawlers, and viruses, Trojans and bots, are constantly trying to gather restricted information or takeover and control connected devices. Because of this every computer and every network connected to the Web must have a functioning Firewall and up-to-date antivirus software.
But it is not well understood that antivirus software will only work against known, widely spread, viruses and Trojans. A competent programmer/hacker can relatively easily write a one-off program that will not be detected by antivirus software but will relay information from a target machine, possibly indefinitely without detection.
There are a few organisations that a Corporation will trust. Chief amongst these is their Bank. But it is unlikely that they would trust even their bank with their core intellectual property, unless it was within a secure safe deposit box in the vault.
Although trust may develop over time, most of the vendors offering services within the Cloud are relative newcomers, without an established reputation for confidentiality. Further, because of the security issues discussed above, it might be difficult, given a leak of sensitive information, to be confident that it was not due to any impropriety by a service provider.
There is also a trans-national issue. Many of the data processing and storage facilities used by the Cloud are located in foreign countries, subject to foreign laws. Foreign Security Services may demand access to such data. Indeed there may be a permanent ‘back door’ to allow monitoring by such Services (as may already exist for some telephone services).
Internet bandwidth is heavily constrained by the currently available technology. This bandwidth is typically more than an order of magnitude less than is available from current, wired, corporate local area networks. Such networks could relatively easily be improved substantially using existing technology. Only cost benefit considerations prevent this being implemented now. But broadband Internet technologies are constrained to a few megabytes per second by available radio spectrum allocations, the high cost of optical fibre distribution, the limitations of present electronic switching systems and the increasing number of users vying for limited resources.
Existing Internet bandwidths are adequate for some applications, particularly when virtualisation is employed. The enabling technology substitutes computing power for bandwidth. Screen images and minimal information are sent to the remote device and the user’s processing commands are sent from the remote device to the server. These technologies work particularly well when all data is held at a central location as is generally the case if the user is operating a handheld device in a remote location.
The problem with this approach arises when large amounts of data are required at the user’s end. If the user is operating a workstation in a workplace in a creative environment they often want to create and locally print large documents, compile software, save media presentations or edit images. In these cases data and needs to be held locally and accessed using a high speed local area network. Using Cloud based or thin client applications and remote processing is problematic.
Cloud base systems add a significant infrastructure overhead compared to local computing. They interpose a significant layer of additional hardware and additional software is required in each case, increasing overall complexity. But these complexities and costs are shared by vastly more consumers and so overall the customers generally see a saving.
The local area network (LAN) has been around 30 years and will live on for some years yet. It provides a secure trusted and relatively inexpensive environment for data intensive, creative, and sensitive computing. Indeed many households with multiple devices are installing network attached storage devices (NAS boxes) and shared printers to create a local LAN that is in turn connected to the Cloud through their ISP.
But the Cloud is now positioned to relace the older Wide Area Networks (WANs).
Communications between distant sites are subject to the same bandwidth limitations as Cloud based computing and costs can be substantial. One way of reducing this cost while retaining a corporate wide area network is to implement some of the features of the Cloud.
It is now often more cost effective to use the public Internet to communicate to a small to medium remote office than to maintain a dedicated private connection. This trend is likely to continue as the public bandwidths expand with increasing fibre rollout. Inter-office communications, including Video Conferencing; database access; and file transfer can all be accomplished with adequate securely without resorting to a conventional WAN architecture. Under this model individual offices will still retain a conventional LAN for network data storage file sharing and printing but each LAN will be an island within the Cloud.
Perhaps the only justification for retaining a WAN in future will be that the remote office in question is indeed ‘a centre of data intensive, creative and/or sensitive computing’.
Most corporations are moving towards web enabling many key functions including databases, e-mail and off course intranets; developing their own internal mini-Cloud. Thus we can expect to see an increased use of virtualisation, a thin client technology and remote computing internally.
The public Cloud will increasingly provide desirable features that cannot be delivered by an enterprise based network.
User demand will grow to provide the kind of services that users can obtain at home or beyond the corporate firewall. For example in the near future it will not be necessary to carry a laptop computer everything needed for everyday computing will be included in a mobile phone-like device.
A number of locations may offer flat screen displays linked to the Cloud that can be called and used by the hand held device to display media or to process office work and communications while travelling. A very wide range of display media and input devices will soon be available.
To utilise these services, a proportion of corporate documents and individual users’ work will need to reside, at least temporally, on the central databases within the Cloud.
Corporations will need to look to ways in which they can split their computing requirements between the internal systems and the Cloud.
At the moment we have two ways of connecting, through our firewalls to the www and through the telephone system. In future these will need to be rationalised.
To optimise corporate capability all of these systems will need to be integrated and services employed where they are available.
One of the principal criteria will be the sensitivity of information and the audiences organisations wish to address. For example, most organizations of information they actively promote to the public at one end of the spectrum and information that is commercially sensitive or secret and the other.
Public information includes product information, corporate positioning, and safety information. This information can beneficially be hosted in the Cloud. It includes: all public websites; extranets; and most Intranets; and some related databases; as well as non-sensitive documents and email.
Secret information includes unprotected trade secrets, confidential board information and commercially sensitive information not yet publicly announced. In government it includes Cabinet-in-Confidence documents and National secrets. This still needs to be kept on a secure server in a secure network. Machines storing top secret information should be physically secure and not permanently connected to a network at all.
Somewhere between these extremes lies a line inside which information must be kept secure and internal and beyond which information could be kept on relatively secure data storage within the Cloud.
We are already familiar with documents passing between networks and between individuals and corporate networks. These documents are passed has attachments to email. We can expect the development of new software systems to facilitate document passing between secure corporate environments and the Cloud.
As these systems develop we can expect to see corporate systems drawing in their boundaries and becoming more specialized; islands of security within a new world of globalised networking.