IT Manager's Handbook: Getting Your New Job Done (37 page)
Read IT Manager's Handbook: Getting Your New Job Done Online
Authors: Bill Holtsnider,Brian D. Jaffe
Tags: #Business & Economics, #Information Management, #Computers, #Information Technology, #Enterprise Applications, #General, #Databases, #Networking
The Many Flavors of Cloud Computing
If you are familiar with any of the Internet e-mail solutions (e.g., Yahoo, Gmail, HotMail) you already have a sense of cloud computing. For these, all you need is a web browser and an Internet connection and you have full access to your e-mail. Everything else you need to be fully functional with these vendors' e-mail solutions exists on the Internet. These solutions have grown to include address books and calendars. Google also has its Google Docs and Google Apps offerings, to include the ability to create, store, edit, and post spreadsheets, presentations, and word processing documents online. Take a moment to note how significant this is. Using Google Apps, a user can do most of the most common productivity tasks without having to purchase any end-user software. Not only does this reduce the software costs, but it also eliminates all the tasks and costs traditionally associated with end-user productivity tools (e.g., deployment, upgrades, patching, licensing, purchasing). It also reduces the burden of IT regarding centralized storage, servers, backup, and so on.
However, it's important to note that cloud computing does not always refer to free solutions. Sometimes it's free, and sometimes it isn't. Generally, cloud computing offerings are free for individual personal use, but come at a cost for business uses. One example is the Google Docs offering mentioned earlier. While all those features and functions are free, Google also offers a premiere edition (Google Apps), for business users at a cost of $50/year/user with additional features and functionality targeted to a business (e.g., more storage, data migration, phone support, shared e-mail distribution lists, greater access control options, shared calendaring). The list of vendors in the cloud computing space is growing dramatically with various offerings from names such as Microsoft, Amazon, Oracle, Rackspace, Salesforce.com, and IBM.
Taking things to the next level are business apps that are available via cloud computing. Best known here is Salesforce.com, which has a suite of business applications for sales, customer service and support, sales analytics, marketing, and so on. These apps are all accessed using only a browser from Internet-connected devices (including smart phones and tablets). All of the usual infrastructure requirements (servers, data centers, management, upgrades, security, backups, databases, etc.) become the responsibility of the provider. Other applications that are making inroads into cloud computing are Human Resources Information Systems (HRIS), payroll, accounting, and content management. Because managing e-mail can be so mission critical, burdensome, and complex, many companies are now offering Microsoft's Exchange e-mail solution via the Internet as a way to shift that burden away from individual IT departments.
As a fairly new industry, cloud computing comes in different flavors, with different terms—some of which overlap:
•
Infrastructure as a Service (
IaaS
)
refers to offerings of servers, storage, Internet connectivity,
firewalls
, routers, switches, etc.
•
Platform as a Service
(PaaS)
refers to a solution that not only provides the hardware infrastructure, but also the software and development platforms for your team to build, install, and maintain applications. PaaS offerings also include version control, testing tools and environments, and ability to roll out and roll back different versions.
•
Software as a Service
(
SaaS
),
software may be browser based or installed on workstations. However, the distinction with SaaS vs. traditional software is that you pay for actual use as opposed to individual licenses. This shifts software to more of a variable operating expense, as opposed to a fixed capital expense. Google Docs and Apps, Salesforce.com, Success Factors, NetSuite, and Microsoft's Office 365 are popular providers of various SaaS offerings. SaaS may not always be for end-users; back-end services (like e-mail anti-spam/anti-virus) can also be an SaaS offering.
•
On-demand computing,
a catchall term that refers to the concept of providing computing resources as they are needed, as opposed to the traditional full-time basis. Also sometimes referred to as “utility computing.” Amazon's Elastic Compute Cloud (EC2) is a popular vendor for on-demand computing.
•
Grid computing
refers to leveraging the processing power of a large number of networked computers for large-scale computations. This usually has the most relevance in the scientific communities. Also sometimes referred to as “distributed computing.”
There can be overlap in some of the terms and models presented here, and definitions may be slightly different among providers. The coauthors of this book used cloud computing—they built a web site on Google Sites—to collaborate on manuscripts.
For some of the pros and cons of cloud computing, see
Table 5.2
.
Table 5.2.
Pros and Cons of Cloud Computing
| PRO | CON |
| Pay only for services you use. | Tied to the vendor. Raises concerns about vendor's viability or issues if you need to change providers, migrate data, etc. |
| No capital outlay for software and/or hardware. No direct costs associated with maintenance, upgrades, support, etc. (All of these are generally bundled in with the service cost.) | Lack of control. You may have little or no say regarding operations, future direction of software, etc. |
| Vendor's dedicated staff for service and support around the clock. Because the vendor is directly vested in the success of this model and their offerings and are able to leverage resources and tools on a grander scale, they may be in a better position to monitor/manage the resources than you could internally. | Resources may be configurable by you, the customer, but not entirely customizable. |
| Shifts a great many of IT traditional responsibilities (deployments, upgrades, patching, monitoring, etc.) to the vendor. | Long-term ongoing costs may approach costs for maintaining in-house resources. |
| As a result of the vendor's focus on reliability, availability, and redundancy, it may often provide built-in business continuity and disaster recovering solutions for your enterprise. | May be complex options to choose from (e.g., dedicated vs. virtual resources), making the cost analysis difficult to evaluate. |
| A variety of offerings giving you a great deal of flexibility. | Could be complex (or impossible) to integrate across multiple on-demand/cloud-computing vendors, systems, applications. |
| Since you're generally just paying for use of the application, which the vendor is monitoring, the concerns of licensing and auditing are almost eliminated. | Concerns about security, privacy, and meeting your organization's compliance and requirements needs. |
| Easy to adjust to changing needs, cycles (e.g., when additional storage is needed, it is available virtually immediately). | With all resources managed and operated by a third party, you have no control over availability and reliability. |
| | Who actually owns the resources, particularly the data? |
Many of the same considerations and factors associated with choosing an outsource provider (see the section
“Outsourcing and Offshoring”
in
Chapter 6, Managing the Money
on
page 182
) go into decisions about computing in the cloud.
Private versus
Public Cloud
The cloud solutions that were described earlier are examples of public clouds—they are there for anyone or any organization that wants to pay. An alternative is a “
private cloud
.” Private clouds are attempts by very large organizations to leverage their significant IT resources to essentially offer on-demand/cloud-like offerings to their internal customers and subsidiaries, as a way to provide services in a more efficient, flexible, and scalable manner.
A hybrid of the public and private clouds is the
virtual private cloud
(
VPC
). The VPC essentially uses the public cloud model, but the cloud provider logically segregates your resources from its other customers using technologies like VLANs and other security solutions so that the resources from the cloud provider seem like an extension of your own network.
Some organizations are more comfortable using a VPC model, since it offers additional security provisions for their initial forays into cloud solutions.
Considerations When Moving to the Cloud
While the cloud solution for IT resources can be very attractive, it is not without potential hurdles and pitfalls. The industry is still very new, and everyone is learning as they go. In time, things will stabilize and become more standardized, but until then make sure you do careful investigation and evaluation before moving to the cloud:
•
Data Retention for Software as a Service (SaaS).
Is your data being kept indefinitely? Or will it be archived/purged after a period of time?
•
Backup for Infrastructure and Platform as a Service (IaaS and PaaS).
If you have an application or database in the cloud, you may be the one responsible to make sure it gets backed up. If you accidentally delete a file, you'll be the one to do the restore. Typically, backups for IaaS and PaaS are done by using additional storage from your cloud vendor.
•
Licensing.
Which software and licenses are you responsible for, and what is your vendor providing?
•
Data location.
In which country will your data be stored? This may be an issue for some regulatory and privacy concerns as well as which encryption software you use.
•
Vendor support.
What sort of Service Level Agreements (SLAs) will your vendor commit to? If there is an extended outage will you get a credit? Is their support team responsive? What kind of tools do they have to help you manage, monitor, and troubleshoot? For SaaS models, will your end-users be calling the cloud provider for support or your IT team?
•
Term of agreement.
Do you have to commit to a multiyear contract upfront? What if you want to get out early? If you do want to leave your cloud provider what about issues of migrating/transitioning data, applications, services back to your organization or to another provider?
•
Organizational.
What impact will it have on your team? Will roles be eliminated or changed? Will new positions be needed to focus on cloud operations and activities?
•
Standardization.
If your environment has unique requirements, or nonstandard/traditional configurations, will you be able to replicate that in the cloud? Most cloud vendors allow you a very defined degree of flexibility—will that be sufficient?
•
Operations tasks.
Who is responsible for installing software? Applying patches and upgrades? Monitoring and management of the resources? These can vary greatly between the IaaS, PaaS, and SaaS models.
•
Billing and invoicing.
Does the provider offer sufficient detail and reporting so that you can fully understand the bills? Which tools are in place for you to ensure that your team isn't using on-demand services needlessly (like the programmer who requests some virtual servers and storage for some testing in an IaaS model, but then forgets to delete them, incurring 10 months of usage fees)? What happens if you are late with a payment—is your service cut off?
•
Liability for breeches.
What responsibility will your provider agree to in case there is a security breech?
•
Recovery/redundancy.
What sort of capabilities/facilities does the provider have in case of failures/outages?
•
Protection of data.
Most cloud vendors consider data security to be a top priority. But how can you be sure they are doing everything they say they are doing, or that you need them to do? For example, if they say that firewalls are configured to prohibit certain types/sources of traffic, will they show you those configurations?
•
Network connectivity.
Do you have sufficient bandwidth at all your locations if your applications are moved to the cloud?
•
Return on Investment (ROI).
It's entirely possible that a long-term cloud-based solution may be more expensive that hosting the solution yourself. However, cloud-computing still offers great advantages in terms of flexibility, scalability, and other conveniences that make it worthwhile. In fact, a 2011
CIO Magazine
survey found that 48 percent of IT leaders are evaluating cloud options before making any new IT investments.
Some of the concerns on this list will vary (or even disappear) depending on the model of cloud service you're using (e.g., SaaS, PaaS, IaaS). However, the list should give you a sense of the issues associated with cloud computing.
Cloud computing is a very new area, and most organizations are just beginning to explore and think about how it can be used. A 2011 industry survey by
Baseline Magazine
shows that the cloud is viewed with many benefits as well as challenges (see
Table 5.3
).