Tag Archives: SaaS

Making Dollars and Sense out of Cloud Computing- An Article by Phillip Beidelman

Posted on October 22, 2012 by WTC Consulting

“Making Dollars and Sense out of Cloud Computing” – how does one do that?  Although here are many examples of economic advantages associated with some cloud-based applications such as email, document sharing, and some forms of data storage, I cannot report experientially that the cloud is always a sure bet for enterprise services based on cost performance.

The popular wisdom says that using the cloud is cheaper, but lower cost turns out not to be the compelling reason for cloud adoption at the enterprise level.  While the expectation
of lower cost may catch everyone’s attention, accelerating delivery of services and allowing us to do things we otherwise cannot do are often the more compelling reasons for adoption.

To make some sense of the current state of cloud computing with an emphasis on the financial aspects, I will establish a context for my comments by first introducing basic concepts and vocabulary peculiar to cloud technologies.  Next, I will report what WTC as a consulting firm has observed in higher education regarding the cloud including what is driving interest in this area.  I will close with comments about the potential role of the cloud within our institutions along with some cautions.

Basic Concepts and Vocabulary

Using a cloud to characterize access to an undefined set of computing resources finds its roots in many early technologies.  I recall in the late 1970s and early 1980s using cloud symbols with telecommunication engineers to denote the idea of large networks using cloud images to show then-emerging virtual private network (VPN) technology.  The underlying cloud concept dates back to the 1950s with large‑scale mainframes.  It was common to share physical computer access from multiple terminals to drive down the cost of cycle time.  The thought was then – as it is today – to depict pooled technology resources as an abstraction of the underlying infrastructure.  The use of a cloud showed where the dividing line was between the provider and the user.  Cloud computing now extends this boundary to cover servers as well as the network infrastructure commonly found in the Internet.

A good contemporary definition is the one used by the Educause Center for Applied Research (ECAR) describing “. . . a style of computing where massively scalable IT‑enabled capabilities are delivered as a service via the Internet.”  (Source = Attributed to Gartner, “Demystifying the Cloud” ECAR Research Bulletin 4, 2010).

The National Institute of Standards and Technology (NIST) goes a bit further and gives us more structure by defining the cloud as “. . . a model for enabling ubiquitous, convenient, on‑demand network access to a shared pool of configurable computing resources (e.g., networks, servers, storage, applications, and services) that can be rapidly provisioned and released with minimal management effort or service provider interaction.”  (Source = NIST Special Publication 800‑145).

So much for definitions. What matters is that all cloud services are not created equal and that NIST establishes that all cloud models share five essential characteristics, have three service models, and four deployment models.  Let’s start with the essential
characteristics.

Essential Characteristics

Cloud services have five characteristics: 1) On‑demand self‑service- you can easily access the service from just about anywhere at any time;  2) Broad network access – access can be both fixed and wired, slow and fast, and again, ubiquitous;  3) Resource pooling – there must be a common set of resources shared by many users; 4) Rapid elasticity – growing and shrinking on demand so that you get what you need when you need it, and  5)
Measured service – ideally you will only pay for what you use.

Service Models

These five characteristics are applied against three broad service models. 1) Software as a service (SaaS), 2) Platform as a service (PaaS), and 3) Infrastructure as a service (IaaS).  Following is a little on each of these.  I acknowledge my generous use of NIST sources in defining these ideas.

Software as a Service (SaaS) is the capability for consumers to use the provider’s applications running on a cloud infrastructure.  Facebook, Salesforce, Hotmail, Gmail, Pandora, and Garmin are examples of SaaS.  The software is delivered over a browser and eliminates the need to install and run applications on the customer’s own computers/servers.  Customers do not manage or control the underlying cloud infrastructure including network, servers, operating systems, storage, or even individual application capabilities, with the possible exception of limited user‑specific application configuration settings.

Platform as a Service (PaaS) is the capability for consumers to deploy consumer‑created
or acquired applications onto the cloud infrastructure and to use programming languages and tools supported by the provider.  Google App Engine, Force.com, Windows Azure, WOLF,  AppFog, and Parse are examples of PaaS.  This approach is primarily for software development to facilitate development and deploy applications without buying, managing and configuring hardware, middleware and software layers.  The consumer does not manage or control the underlying cloud infrastructure including network, servers, operating systems, or storage, but does have control over the deployed applications and possibly application hosting environment configurations.

Infrastructure as a Service
(IaaS) is the capability for consumers to provision processing, storage, networks, and other fundamental computing resources including operating systems and applications.  Amazon, Rackspace, GoGrid, CloudSigman, and Nervanix are examples of service providers offering IaaS.  The consumer does not manage or control the underlying cloud infrastructure, but has control over operating systems, storage, deployed applications, and possibly limited control of select networking components such as host firewalls.  The computing infrastructure is typically billed on a utility basis with the amount of resources consumed usually reflecting the level of activity.

Deployment Models

Finally, there are four deployment models: 1) Private ‑ operated solely for an organization; 2) Community ‑provisioned for exclusive use by a specific community of consumers with shared concerns; 3) Public ‑ available to the general public; and 4) Hybrid ‑ composition of two or more clouds bound together by standardized or proprietary technology that enables data and application portability.

The idea of private cloud deployment models turns out to be a big deal in higher education with the developments in I2 demonstrating the potential for services such as IP Centrex
offered by CENIC in California.

What WTC Sees Going On

Tactically: There is no question that cloud services are off and running, but our experience leads us to believe the actual rate of adoption is most likely slower than the market and some surveys might have us believe.  Here are some facts to back up that idea.

Public cloud workloads have had a 50% Compounded Annual Growth Rate (CAGR) over the last 3 years and server growth has shifted to the cloud with server shipments into public clouds expected to grow at a 60% CAGR through 2013.  On‑premise server growth is expected to decline.  (Source = Morgan Stanley Research May 23, 2011, Cloud Computing Takes off Global survey of IT managers on cloud migration).  Curiously, in that
same study only 8.6% of cloud users cited server hardware as the primary area of savings from the migration to cloud computing and only 1% cited decrease in storage spending.  No compelling financial argument there.

Meanwhile in higher education, there is a different story.  A tracking poll of 150 schools conducted by CDW-G in 2011 shows 29% of higher education institutions have written strategic plans for cloud computing and institutions expect to spend 15% of their IT budgets on the cloud within two years.  Additional information from that poll shows
cloud adoption in higher education to be in the following phases:  32% in discovery; 29% in planning; 28% in implementation; 6% in maintenance, and 5% not considering anything.  Our experience with client engagements that examine applications run in large higher education IT organizations and the full cost of IT services does not match the CDW-G tracking poll.  While we recognize it is possible our clients simply do not mirror the poll, we believe that what is actually happening is a pursuit of the easier targets because they are either completely or virtually free.  This situation is borne out from other results showing that Gmail and Google Docs are both reported as adopted by 50% of respondents, MS Office Live Meeting by 22%, and Salesforce by only 2%.  18% of
respondents indicated no adoption.  We believe this means that once free services are off the table, the rate of adoption in higher education will be slower.

Strategically: So, what is pushing the cloud?  At a technical level, convergence improvements have made the cloud possible.  Leading the pack of key innovations are 1)
virtualization software to run multiple organizations on common physical infrastructure as if independent, 2) the eXtensible Markup Language (XML) to have a standard way to interchange data between providers and users, 3) Web 2.0 technologies to allow web‑based applications to have better user interfaces using standards‑compliant browsers, and 4) cheap commodity internet bandwidth reducing the cost of access.  But the real story is just emerging.

What we see as the big drivers for cloud adoption at an enterprise level will be accelerated-time-to-market for new services, increased access to resources, improved functionality, and getting more done with less.  We think cost reduction will be a secondary motivation.

Compelling evidence for this point of view is found in a set of statistics published by IBM regarding one of their PaaS offerings summarizing experiences of 2,000 of their engagements in the first 6 months of 2011 with 4.5 million daily client transactions, and 1
million managed virtual machines.  The results look to us as a portent for the real message of the cloud.

If the IBM results are anywhere near correct, then we can see the future now.
These results are not about saving money directly, rather they are about improving operations and potentially bringing teaching and learning and research to new levels not formerly available.

Where the Cloud Fits in and Some Cautions

Where to look. So we know the various flavors of cloud services available and can get enthusiastic about their potential role.  Now the question is where do they fit.  It is
useful to acknowledge that not all applications are a good fit for the cloud.  At this point in cloud development, the decision to consider the cloud is driven by a few things: 1) the importance of the application, 2) institutional risk, and 3) economics.  We know that most IT organizations use about 400 different applications, not counting new applications being considered, those on a wish list, or anything involved with special research needs.  Following is our suggestion about how to determine which applications might be cloud
candidates.

First, consider the level of sensitivity associated with the data.  The higher the sensitivity the less likely the cloud is suitable because it is difficult to ensure treatment of data when you do not know where it is and who might have access to it.  Second, consider the degree of importance of the data.  If data is critical to the institution, then again the cloud is a
questionable place for it.  Third, if the operating system you need is not available, then your options are limited.  Fourth, since many pricing mechanisms are data and usage sensitive, applications requiring large uploads or down loads may not be a good choice.  The fifth consideration is licensing.  Some license schemes do not move well to the cloud.
Finally, if data retention is important, then the cloud may not be a good choice.  In summary, filters to determine cloud candidacy would be sensitivity, criticality, OS restrictions, frequency of access and file up and down load size, licensing, and data
retention.

Some Cautions. The advantages of the cloud also have some gotchas.  We suggest you consider some of the trickier decision points: 1) Ownership – how do you secure ownership, 2) Governing law – which state law governs (e.g., where you are, where the provider is, or where the data is), 3) Service Level Agreements – can you really make one that is enforceable, 4) Failure – what to do if and when things go wrong, 5) Disaster Recovery & Business Continuity – what happens when your side or the provider side of the technology fails?  6) Single Point of Failure – is the cloud provider really just that?
When you are ready for a full list of things to think about, you will find “Above the Clouds: A Berkeley View of Cloud Computing” Feb 10, 2009 a useful document.

Summary

The cloud has a solid foothold in higher education in commodity services such as email.  The continued expanded use of the I2 infrastructure will provide a national and global private cloud capable of being used effectively in teaching and learning and research.  Having said that, caution is still a good thing when it comes to enterprise IT.  We believe the rate of adoption will be slower and less financially motivated than some would think.  But the next five years will tell the tale.

 

This article was published in the Fall 2012 ACUTA Journal Volume 16, No.3.

Posted in About WTC, Trends | Tagged , , , , , , , , | Leave a comment