Tekexpertise Blog

Cloud Computing has matured greatly in the last year and is now a major technology shift. Basically, Cloud Computing means running applications across the Internet. In some senses the “cloud” is just a metaphor for the Internet itself. Organizations outsource applications to vendors who host these applications in their own data centers. Organizations access them over the Internet using a web browser.

With Cloud Computing, instead of owning a specific server or software license for an application, an organization pays a monthly fee. Details of hardware and software provisioning, configuration, management, monitoring, disaster recovery, etc. are provided via the Cloud vendor thereby decreasing the administrative and technical burden on organizations.

Cloud Computing has become feasible by the latest increases in Internet bandwidth, storage capacity and server speed as well as advances in software and tools. As each of these becomes ever more powerful, the value proposition of Cloud Computing becomes more compelling.

Major vendors such as Microsoft, Amazon, Google and others now have Cloud offerings. These companies have the capability to massively scale data centers to accommodate the demands of customers. Also, almost all organizations cannot provide an environment that is nearly as bullet-proof as these vendors can provide. The recent competition between HP and Dell over 3Par shows that large vendors are placing big bets on Cloud Computing as a new IT paradigm.

A major inducement to Cloud Computing is that the start-up costs are much smaller than provisioning the same application in house. However, it is not necessarily true that Cloud Computing is less expensive in the long term. An analysis of the economics of in-house IT infrastructure versus a Cloud implementation needs to consider not only the initial purchase of the hardware and software, but also the hidden costs of the on-premises power and cooling, the costs of the staff to properly maintain the infrastructure, software licensing costs, etc.  Furthermore, it is likely that as more competition comes to Cloud Computing as well as continuing improvements in computing overall, that costs will decline or remain steady while the value of the offerings continue to improve.

All organizations should now begin serious efforts to move at least some of the IT infrastructure to the Cloud. If you would like to discuss this in more depth or if you would like to put together a plan, please send an email to ed.mchugh@tekexpertise.com.

Many organizations have postponed server upgrades during the recession. As the economy recovers, these organizations may now be thinking about upgrades. Their servers have may be getting beyond useful life and are unable to run the latest or supported versions of software. Of a greater concern is whether their servers may be at risk of hardware failure.

Servers today are cheaper and far more capable than those bought four or five years ago. A good rule of thumb is to replace servers no sooner than three years after the original purchase (they are still under manufacturer’s warranty) and no later than five years. After five years, the probability of hardware failure, particularly for disk subsystems, goes up noticeably.

However, the best upgrade approach may not be replacing existing physical servers with the latest server hardware. There are two strategies to be considered before making that purchase. First, if there is more than one server that needs replacing, organizations should consider buying a more powerful server and use virtualization to replace two or three (or perhaps more) physical servers. Virtualization is the technique where multiple instances of server software, including the operating system, can be run on the same physical hardware.

An even more intriguing possibility is to retire at least some physical servers completely and outsource their functions. For example, organizations should not be hosting their own websites internally. Other prime candidates for outsourcing include email, CRM and Intranet servers.  Unless there are compelling reasons, there is little advantage to keeping those functions on internal servers.

For a relatively modest monthly fee, organizations can remove the administrative burden and expense of running their own servers. Further, these hosting companies provide the redundant power, Internet connectivity, spam and virus prevention, monitoring and other functions that most organizations would not be able to do cost-effectively themselves.

The number of hosted solutions is continuing to grow as “Cloud Computing” matures. I will be discussing Cloud Computing in a future blog post.

If you would like to discuss this in more depth or if you would like to put together a plan, please send an email to ed.mchugh@tekexpertise.com.

Every organization needs to be prepared for adverse events that could disrupt normal operations. Business Continuity Planning (BCP) is the discipline where organizations plan to handle these events.

A BCP strategy must have the flexibility to recover from partial outages as well as complete outages. Most events are localized and two of the most common are power failures and loss of Internet connectivity. Also, outages can include weather-related events such as snowstorms where employee safety requires them not to come to work.

The BCP process begins by analyzing the key business activities and their operational requirements. Interdependencies between activities need to be determined as well. Further analysis should reveal how long an interruption can last before the impact becomes unacceptable. Note that organizations can’t react too quickly to an event that may be of a short duration while waiting too long to implement the plan would cause valuable recovery time to be lost.

The analysis proves the information needed to create a plan. This plan has detailed information how each business activity will be recovered and where the recovery will take place. Recovery locations could include alternate sites, work from home or an unaffected part of the main site. The plan also should address the order in which each business activity is restored and when and how activities return to the affected site.

No plan is adequate unless your staff is trained on what is in the plan and what their roles and responsibilities are in case of an outage. It is also critical that run-throughs of the plan be done on a regular basis. New employees need to be trained on the plan when they join the company.

One of the key items of your plan is to identify one or more incident managers who have the authority to ensure that the outage is handled appropriately. An incident manager must have a clear mandate and the ability to direct people in all departments who are affected by the outage.

No plan is complete unless there is an ability to reach all the effected personnel to give them information wherever they might be. Automated systems are available that leverage multiple means of communication (home phones, cell phones, texting, etc.).

If you would like to discuss this in more depth or if you would like to put together a plan, please send an email to ed.mchugh@tekexpertise.com.