Monthly Archives: December 2013

Mitigating Data Center Service Models: IaaS, PaaS, SaaS

On April 21, 2011, Amazon’s Elastic Compute Cloud (EC2) experienced a major outage at its data center in Virginia that brought down important websites including Foursquare, Springpad, Reddit, Quora, BigDoor and Hootsuite. The outage lasted from 5:00 am until the late afternoon that Thursday. However, many problems were not resolved until April 25. The EC2 crash made headlines worldwide because of the high number of important websites affected. However, it also served as both a wake-up call and a textbook example from which data center managers should learn from and prepare as the demand for data center continues to increase.

Data Center Service Models - IaaS PaaS SaaS
Growth spawns increased risk. The data center industry is not immune from this fact as the Amazon EC2 example shows. However, risk can be managed, and future harm mitigated if the correct computing service delivery system is implemented. There are three categories: Software as a Service (SaaS), Platform as a Service (PaaS) and Infrastructure as a Service (IaaS). Each is different, and each has its strengths and weakness.

Software as a Service (SaaS)

SaaS is computing solution service that allows software to be deployed over the internet. The major advantages of SaaS are that it is very often user-friendly, and it allows internet access to popular commercial software. Software is usually managed in one central location. This central management relieves the end-user of the hassle of having to deal with software upgrades, patches, and dealing with application programming interfaces (APIs).

The biggest disadvantage is that SaaS may not be adequate when an end-user requires extremely fast processing of real time data. Also, in cases where end-users need continuous modifications to their software and solution needs. There might also be concerns by end-users who are prohibited by law from hosting data externally.

Platform as a Service (PaaS)

PaaS combines the benefits of SaaS but includes the added benefit of allowing software development. Not only is the software delivered over the internet, but end-users are given access to a platform for creating software.

PaaS is perfect for end-users who need to have the ability of deploying software over the internet, but that also want a platform where a team can collaborate to develop or improve its existing software.

The biggest disadvantage of PaaS is that certain end-users that develop software will not want to disclose or give access proprietary software language. Moreover, some end-users need applications that are highly portable or that will require changes of the underlying data center’s hardware and software. In these cases, PaaS might be inappropriate.

Infrastructure as a Service (IaaS)

IaaS offers a complete suite of on demand services to end-users including access to servers, storage, network and operating systems. It is the combined benefits SaaS and PaaS with an actual system.

Some data center offer “public” IaaS where this computing infrastructure is offered over the internet to multiple end-users at a time. However, come data centers have a “private” IaaS that serve a private network for one end-user.

IaaS is perfect for end-users that have recently gone into business and do not have the funds necessary to invest in computing infrastructure. It is also good for end-users that are experiencing rapid growth but that are uncertain if a capital investment is worth it as well as in cases where an enterprise needs to limit computing expenses.

The biggest disadvantage of having an IaaS data center service model, whether private or public, is that there are constant regulatory changes that may prohibit end-users from outsourcing or storing data externally. One day your IaaS maybe flourishing, the next it has dried up because the end-users regulatory requires that store all of its data domestically.