[Part 1 reviews the recent history of cloud computing, including virtualization, what drives the cloud, private vs. public cloud computing, and the likely dominant players.]
THE CURRENT STATE OF CLOUD COMPUTING
Companies of all sizes are using the public cloud and many are driven to it by services like Salesforce.com. Salesforce.com has a satisfied end-user base, which drives its usage much more than the fact that it is a cloud application.
This indicates that cloud usage is driven by the innovative applications it provides and much less by the technical aspects of the cloud. This makes sense, as end users just want the solution an application provides, and they want satisfaction now rather than being put in a long IT backlog. As with all things technical, the applications drive the platform and each new platform enables new applications Ė the two go hand-in-hand.
Cloud usage Patterns
There is a trend to setup private clouds, which provide the upside of the public cloud but with fewer risks and possibly lower costs. With private clouds, IT regains control, and with the right front-end tools to create services on demand, IT is able to act like a public cloud vendor in the eyes of the end user while at the same time assuring account control, data management, and the quality of services provided.
This is a very large win for both parties, and it is likely to be a solid choice for mid-to-large companies for a long time as they can afford to host and manage a modern private cloud. Even small firms can quickly create private clouds with products like those from VMware (www.vmware.com/solutions/cloud-computing).
It is important to note that a private cloud is different than a traditional data center. A data center hosts servers in a secure location like a private cloud, but the private cloud adds end-user front-ends where services can be purchased with pushbutton simplicity. More vendors will add support for private clouds in this area, which will greatly increase the value of private clouds.
Hosted applications such as Salesforce.com are one area that private clouds do not cover as software application developers do not allow hosting of their applications. This means that private clouds are good for custom applications and raw servers-on-demand only and a combination of public and private clouds are likely to being used.
Who Will host the Cloud?
There are two cloud hosting possibilities in the future. One possibility is that a few massive providers will host most cloud services and applications. The other is that there will be many hosting providers, each focused on a niche such as health, security management, server message block (SMB), large The Current State of Cloud Computing 329 federal government agencies, and so on.
As with all markets, the evolution will start with many providers, and after a period of time, the strong will rise and the others will fall back to a small niche. Over time, a combination of a few large providers such as Microsoft will host thousands of small applications, and many small providers will hold a strong niche use to a strong offering that appeals to a specific group of users.
Cloud Computing and Collective Intelligence
In addition to low cost and convenience, a third key element of the cloud is collective intelligence. This is the ability of cloud applications to know what everyone in the community is doing and using this information to make the community work better. A well-known example is Amazon, where users can quickly determine if a book is well liked by many people or not liked at all, enabling a quick buying decision. Another Amazon feature is the ability to buy the same things a user just like you is buying on the assumption you will like it too. This can be very powerful once a user base is established and understood.
Collective intelligence can be used in any type of online community. For example, in the security world, we could quickly determine if a new security patch is safe to install and at the same time know what the risks are if the patch is not installed. Using traditional methods, if a vendor releases a security patch for a Web server, an IT person needs to determine if the risk of the patch breaking the Web server and taking down a business application is higher than the risk of waiting to install the patch at the next maintenance window. This is a very tough call to make in a 7 ◊ 24 environment.
With a cloud-based solution running with a well-populated database, an IT person can quickly determine if people with a environment similar to theirs are able to successfully install the patch. And for similar users who have not installed the patch, they can determine what level of risk they are at for a security breach. Without the cloud, the IT person could post a question on a trusted user forum site and may get an answer, but the answer cannot be verified. Contrast this with the cloud, where the data is looked at by thousands or (some day) millions of users to find the answer directly from the source.
Another example is a system in which a user can ask "If I add 4 Gb to my server, running the loads I am running, what will be the performance gains?" In a fully-populated collective intelligence system, a number of computers with the same configuration as the one in question will be examined both with and without the additional 4 Gb to determine the gain. Thus, the person asking the question gets a precise answer in a few seconds without the need for trial and error, extensive resource, or use of an expensive consultant. The 330 chaPter 12 The Future of Cloud Computing user can then ask "What if I add 16 Gb?" and review the results again until they have the optimized number for their budget and needs.
There are countless examples of how this technology could be used today if it existed, and it will surely be used in the future.
Over time, collective intelligence is likely to prove to be the key advantage of the cloud. In many aspects, it enables us to achieve the early promise of artificial intelligence, which had a hard time gathering all the data it needed to make real-time and informed decisions. Collective intelligence solves that problem. However, this is still in its infancy because with the exception of retail sites, most current cloud applications do not take full advantage of collective intelligence.