In many organizations, IT departments are no longer in control of the applications or delivery delivered to users. The cloud has activated a new application delivery mechanism in SaaS (software-as-a-service) and its increasing popularity often has IT on the outside looking through storm clouds
Business and functional leaders are directly specifying and buying SaaS solutions like Salesforce.com (sales), Marketo (marketing), SuccessFactors (human resources), or others. Previously, IT would provision datacenters and network infrastructure to deliver applications to their users, and would have complete control and authority over it all. They would also be obliged to assure great availability and response times of those apps and would monitor them with instrumentation often located in the datacenter. With SaaS, IT may not even know of the service purchase, but soon enough, they inherit the burden of diagnosing issues that arise from its use.
The situation will only get worse. Current estimates suggest SaaS applications are a $16B market worldwide and growing more than 15 percent annually, a rate that is 2.5 times the rate of "hard copy" software. Venture capital and acquisition pace is feverish for companies with SaaS offerings serving a wide variety of industries. In the last year alone, there were 300 deals delivering $3.3 billion in funding at valuations averaging 3.5 times sales (most notable are Evernote and GitHub).
In addition, there were more than 200 acquisitions valued at more than $19 billion with six to 10 times average valuations (notables are BuddyMedia and SuccessFactors). SaaS IPOs in 2012 delivered an estimated 43 percent stock appreciation versus the 16 percent NASDAQ mark. This indicates that the way we all buy and use computer applications has and will continue to shift dramatically toward services.
The issues faced by IT and network engineers in this new SaaS-laden environment are many, varied, and have changed greatly from challenges in the past. So much of the application delivery to users is out of the direct control of IT, making it difficult to know where to start or who to blame when issues do arise.
- Users may be outside the walls of a company in unknown locations or on unknown devices. Is it the device, the client-side app, or browser, cellular network, the SaaS provider, or other?
- Users may be working on their personal mobile device (BYOD). Is it the device, the WiFi network, the WAN, the SaaS provider, the VPN (virtual private network), or other?
- Users may be working via PC right at their desk. Is it the client-platform compatibility, a bandwidth issue, an out of date security certificate, the SaaS provider, or other?
- Service suppliers may use a CND (content distribution network) to improve performance of their app. Is the user’s network configured to provide access to the Internet through a common datacenter, thereby significantly defeating the CDN value and performance?
SaaS providers may offer SLAs (service-level agreements), but often they don't. Users, whether through IT or directly as SaaS buyers, had better be checking to see if those SLAs are being met, or monitoring performance closely especially if no SLAs are offered.
Service availability and response time (performance) are most critical to user experience and therefore the measures most important to monitor. These measures should be taken from the point of view of the users themselves, taking into consideration the sites from where access is most often made, times of day, homogeneous groups of users, whether or not a smartphones are in use, and other details of the SaaS accesses. The measures should then be taken all the way through to the real SaaS applications in question. Everyone should be watching their SaaS.