CRM Software Quotes
CRM search»Cloud Computing»A Cloudy Forecast

 Rick Cook Cloud Computing: A Layered Approach to Remotely Delivered IT Services

4 stars Average rating: 4 (from 71 votes)
 By Rick Cook

Waiting Eagerly For The Clouds To Roll In

Cloud computing may be the most eagerly awaited and most touted IT development of the decade. But while enterprises look to put everything from business software applications to storage out the cloud, and while there's a lot of excitement and experimentation, except for some specialized instances like software as a service, most IT staff aren't willing to commit a major part of their operations or IT budget to the cloud just yet. As a result there's a lot more talk than action as potential users wait for cloud vendors to get the final kinks out.

This isn't to say that the first wisps of the cloud haven't arrived. According to Forbes magazine, Software as a service (SaaS) already had 3 million users by 2010, Platform as a Service (PaaS) has over 60,000 applications and there are a whopping 15 million users of Infrastructure as a Service.

Still, compared to the potential market – which is nearly every enterprise in the world that uses computers – these are just tiny numbers and the number of companies using the cloud for mission critical purposes is a small fraction of that.

However the adoption rate is growing rapidly. According to both the CRM Online Guide and Gartner, the 2009 worldwide market for cloud services was $56.3 billion worldwide, up 21.3% from 2008 and expected to grow to $150.1 billion by 2013. Clearly, in spite of the confusion and hesitancy, cloud computing is a major trend in information technology.

As is usually the case, the aggregate numbers conceal as much as they reveal. To understand where cloud computing is going and where the opportunities are today, you have to examine the situation more closely. When you do that a lot of the hype evaporates and you're left with a highly segmented market which is growing at very different rates, and with each segment facing its own challenges.

What Do You Mean "Cloud Computing"?

In fact cloud computing and closely related fields like hosting segment into a number of different offerings, each with its own history, value proposition and challenges. Taken holistically, cloud computing's multiple layers are marked by unique target markets and varying solutions to address different types of business problems.

Networking vendor Cisco breaks the cloud continuum into five segments, some of which aren't usually considered as part of the cloud environment. However all of them involve the fundamental principle of the cloud: Running IT tasks on someone else's computers, usually on a pay as you go basis. Cicso lists the parts of the "cloud chain" as:

Co-Location – Running on someone else's servers or computing infrastructure in a remote location.

Hosting – Having a third party run some of your IT tasks for you on their computers. Most commonly seen with Web servers because they are specialized applications with special needs – notably reliability.

Infrastructure as a Service (IaaS) – Using the vendor's facilities for computing cycles and storage, most often on one or more virtual machines with associated storage and perhaps basic software like antivirus provided by the vendor.

Platform as a Service (SaaS) – A complete ready-to-run framework or environment from the vendor. Most commonly found today in application development, especially web application development.

Software as a Service (SaaS) – The vendor supplies the application as well as the hardware and infrastructure delivery resources. The most common SaaS applications are customer relationship management (CRM) and human capital management (HCM), although there are hundreds of types of SaaS business applications.

Each of these cloud computing segments is in very different stages of acceptance in the market. Aside from hosting and co-location, SaaS and some parts of IaaS, notably storage, have gained the most acceptance. PaaS is well regarded in web development but only getting started outside the software development community.

Cloud Computing Segment Creep

It's important to note that none of these cloud categories are separate technologies. All of them are variations on a common theme and there can be a lot of overlap. What's more service providers tend to expand into other categories as their business models evolve.

For example, Amazon started offering cloud services in March 2006 with S3 IaaS (storage only) as a way to utilize excess capacity in their server farms. In December 2007, Amazon introduced SimpleDB, a non-relational database that allows users to run queries on structured data in real time. More recently Amazon has rolled out Elastic Compute Cloud, a more generalized IaaS, that creeps into the PaaS market, and offers virtual servers for rent and supports Windows, Linux and OpenSolaris. Today, customers leverage Amazon's cloud solutions in varying ways and for individual objectives, such as disaster recovery, virtual desktop infrastructure and handling peak demand for computing resources.

Salesforce.com developed an on-demand CRM system and then leveraged that SaaS experience to later deliver the Force.com PaaS solution. Other major vendors are following similar strategies by starting in an area they know well and expanding their offerings as they become more comfortable with the cloud.

This kind of cross-fertilization makes sense when you consider the highly concentrated nature of the server market. According to Rich Rashid of Microsoft, 20% of the servers in the world are sold to just a few companies: Notably, Google, Amazon, Microsoft and Yahoo!. These companies have tens of thousands of servers in data centers around the world. The second and third tier companies also operate very large numbers of servers, with the numbers gradually tapering off until you get to companies with only one or two servers.

Given the nature of servers, the increasing size of "chunks" of server capacity as servers become ever more powerful, and the declining cost of storage, there are a number of companies with tremendously over-provisioned server capacity. It makes sense for these companies with excess server capacity to find ways to spread that capacity to other companies and recoup some of their IT costs.

Meanwhile at the lower end of the market, capacity management is becoming increasingly tricky. It's hard to "right size" dynamic server capacity and each server represents a significant capital expense to companies with less than 10 servers. What these companies want to do is purchase their computing capacity in smaller increments with the ability to expand quickly as their needs rise, all while keeping a minimum expenditure tied up in data center resources.

That translates into a cloud computing opportunity whereby IT resources can be leveraged for the benefit of vendors and customers. Customers want to buy services as they need them and vendors want to utilize their excess capacity by selling it short term to others.

What all this adds up to is a somewhat confusing, rapidly evolving picture. Vendors are scrambling to meet the perceived needs of potential customers, customers are cautiously dipping their toes in the water and no one is quite sure how it is going to shake out.

The Customer Perspective

When considering moving into the cloud it's important to ignore the hype and concentrate on the real world capabilities and measurable benefits. Among other things, this means looking beyond the cloud's potential and product announcements and aligning what is available and stable today with your organization's business challenges and opportunities.

This is particularly important in the areas of IaaS and PaaS where many of the broadest products, such as Microsoft's Windows Azure and IBM's Blue Cloud are still in the early stages of roll out.

Increasingly the devil in cloud computing is in the details. When considering a cloud service, particularly PaaS or IaaS, it's important to nail everything down with clear service level agreements and other contract conditions.

Utility Pricing

One of the most important – and trickiest – things to nail down is pricing. Unlike SaaS, the pricing on IaaS and PaaS tends to be highly variable, not only from vendor to vendor, but within each vendor as well. Most vendors offer some variety of capacity based pricing, but with so many variations that each contract becomes almost a custom arrangement.

In fact it's possible to use cloud computing today for little or nothing if your usage pattern is right. This is because many vendors have very low minimum prices to encourage people to try their services. For example, Amazon SimpleDB offers the first 25 machine hours, 1 GB of data transfer and 1GB of storage for free each month. Theoretically a small user could use SimpleDB over an extended period at no cost.

The other factor is that many vendors bill for their services by function. For example Amazon's S3 storage service costs $0.15 per GB per month for storage alone. However there are also data transfer charges both in and out. Data transfer in costs $0.1 and data out costs $0.17 per GB. There are also additional charges, for example $0.01 per thousand requests and discounts for very high data usage. Prices are higher in Europe.

Similarly Amazon's EC2 PaaS offers three models. A small instance server comes with 1.7 GB of memory and 160 GB of storage, A large instance server has 7.5 GB of memory and 850 GB of storage. An extra large instance server has 15 GB of memory and 1690 GB of storage. There are more variants here as well.

Other PaaS and IaaS Vendors offer Similar Structures.

What this all comes down to is that anyone considering using PaaS and IaaS needs to know very clearly what their requirements and usage patterns are. Especially for larger applications, the exact nature of your use can cause a major difference in your monthly bill. Not to mention which service offers the better deal.

It's also important to understand that generally you purchase cloud services on a period basis, normally month to month. This is seen as a major benefit of cloud computing, but it does mean that you need to keep careful watch on your usage patterns or you can find yourself with an unusually high bill if something goes out of whack in your data center.

Finally, provisioning is closely related to pricing. A PaaS or IaaS vendor should be able to offer fast deployment as needed by the customer. The usual phrase in the industry is "capacity on demand", however the contract should spell out how quickly the resources will be available. For some vendors, capacity on demand means exactly that. You need it, you self provision and you've got it instantly. Others take anywhere from a few minutes to several hours to get additional capacity to their customers. Obviously the quicker the better, but sometimes that should be balanced against cost. If there's a significant difference between "immediate" and "in an hour" you may want to consider whether "immediate" justifies the price premium.

Stability

Another important requirement for cloud computing of any sort is stability, both long term and short term. In other words, will the service be up without interruption and does the vendor have the staying power to last in what is an increasingly competitive market?

An important consideration for short-term reliability is provisioning across multiple data centers so a single incident can't disable the service. Here again, there is a lot of variation on how vendors deal with reliability issues. Google and Amazon replicate each application instance to multiple data centers in different physical locations.

AT&T is offering Synaptic Hosting designed for handling peak loads. The service, which supports Windows server and several flavors of Linux, as well as Web services, spreads applications between AT&T data centers and charges on the basis of average load per month. AT&T says the service is particularly well suited to computing environments with very high peak loads, such as retailers during the holiday season. Here again, contractual guarantees are important to make sure the vendor will support the service when it is needed.

A special case of reliability is business continuity and disaster recovery. In case of trouble, from a fire to an earthquake or hurricane, the data center image, or critical parts of it, are shifted to the cloud. This can be done in a matter of minutes.

One of the problems with disaster recovery in the cloud is that many businesses may be affected by the same disaster and be looking to shift into the cloud. This can result in bandwidth and other congestion and degrade performance. Some vendors offer premium services at an extra fee that give companies priority in bandwidth and other resources.

Business stability requires due diligence on the part of the customer. You need to make sure that your vendor is going to be around for the long haul. This is especially true in the case of a long-term relationship where business critical information or applications are going to live in the cloud for some time. Customer references, independent audits and financial viability must all be ensured.

Cloud computing is evolving rapidly and as it does there will undoubtedly be shakeouts. Businesses that lived through the wave of consolidation of internet service providers at the beginning of the last decade remember the problems and frustrations as smaller ISPs were gobbled up by larger ones or disappeared entirely. It's important to consider whether your cloud provider will last. This is somewhat mitigated by the lack of long-term contracts in the cloud business, but it still needs to be considered carefully.

Information Security

Of course security is, or should be, a given. Customers considering cloud computing will find that information security is a critical diligence factor when choosing a vendor, and varies dramatically by vendors.

Normally in the case of IaaS and PaaS, the provided environment will include basic security services, such as firewalls and antivirus software. It's important to know what you're getting and consider whether that's enough for the value of the business data you're trusting to the cloud. You will probably want to consider additional measures and you need to be sure that the vendor will allow you to add your preferred security tools.

One particularly important feature to consider is logging and log analysis programs. The vendor will probably offer logging and some log analysis tools, but you may want to add more powerful and complex capabilities to keep abreast of what is happening in your little section of the cloud.

Moving Into The Cloud

Although utilizing PaaS or IaaS today requires a good deal of study, the good news is that things are getting much easier rapidly. As more vendors get the kinks out of their offerings and the business shakes out and settles down, it will get easier to move things into the cloud.

Further, as more businesses become comfortable with PaaS and IaaS and entrust more of their business functions to them, it will be easier to convince corporate decision makers to grow their cloud strategy. Already the potential cost savings, especially in capital expenses, are compelling. Increasingly the move to the cloud is a matter of working out the details. End

How would you rate this article?   

Call Centers
 Filed In Categories: Cloud Computing
Categories
 Tags Tags: Cloud Computing Market
Tags
Trackback Permalink: www.crmsearch.com/cloudcomputinglayers.php
Trackback
Author  Author: Rick Cook
Author
 Share Share:    Bookmark and Share
Share

 

Comments (6) — Comments for this page are closed —

Guest Alexa Monaic
  Is cost savings the biggest driver for cloud computing?
  Denise Denise Holland
    Cost is certainly one significant factor and cost savings can be achieved in terms of initial up front capital expenditure costs, reduced implementation time and costs, lower IT labor costs and often lower total cost of ownership (TCO) over the life of the IT service. At a Gartner conference, the analyst firm indicated the top four drivers of cloud computing are budgetary constraints, resource constraints, time to market and a desire for more control (i.e. permitting business to drive technology, not the other way around).

Guest Laureen McKnight
  What's the difference between hosted CRM and on demand?
  Denise Denise Holland
    Hosting vendors remotely host the application as a service, however, the hosting is generally single-tenant and the service may still require the customer to procure an up front software license and traditional implementation fees. Hosting providers deploy unique software instances for each customer. Hosting is really just outsourcing the hardware and software management to an outside firm. In contrast, on-demand providers leverage a single application (via a multi-tenant architecture) and hardware platform to support many customers thereby achieving greater economies of scale and relieving the customer from software maintenance and upgrades. On-demand pricing is generally much less and based on a subscription instead of an up-front purchase with annual maintenance.

Guest Laureen McKnight
  I'm not sure I'm understanding why anybody would go the hosting route.
  Denise Denise Holland
    Hosts deliver advantages in the areas of implementation, expertise and software control. In a privately managed hosting arrangement, the vendor provides most or all operational support for the CRM application at their site. Because the host has an existing infrastructure, they can usually accelerate the implementation period (not as fast as SaaS, but faster than an on-premise environment). The host employs experts who implement and manage information systems full-time, so they can bring a wealth of knowledge and experience to their clients. The customer preserves the flexibility of being the only company to use the application, and can therefore modify, integrate, customize or otherwise manage the application at will and with a high degree of control. The biggest disadvantage of hosting is cost. Using host resources for the software management, integration, customization, upgrades and the like is in line with hiring consultants, and more expensive than using internal resources.
 

 

Share This Article

 

 

Quote

Software as a service (SaaS) already had 3 million users by 2010, Platform as a Service (PaaS) has over 60,000 applications and there are a whopping 15 million users of Infrastructure as a Service.

~ Forbes Magazine

 

Related Articles

 

More Articles By Rick

 

 

CRM Quote

Follow Us
social
social
social
social

crm search

Home   |  CRM  |  Sales  |  Marketing  |  Service  |  Call Centers  |  Channels  |  Resources  |  Blog