- Traditional network SLA metrics do not take into account changing IT needs.
- Is your network vendor willing to extend service guarantees to application availability?
There are many ways of looking at network service level agreements (SLAs). For telecom providers and certain clients, they can be a mere commercial agreement whereby network downtime will be compensated. In other cases (for example, when downtime can prove very costly or even disastrous to a business), the enterprise customer will need to pay for extra resiliency in the form of five-nines availability or even 100% availability based on 1+1 back-up and/or a 3G wireless broadband data link. Traditional data WAN SLAs still contain the standard metrics, such as jitter, roundtrip delay, latency, availability and MTTR, and this is a good thing overall for making sure the carrier is accountable for the networks. However, IT managers should also be exploring SLAs all the way to applications running on the desktop.
As enterprises put more applications, data and computing power in cloud environments, how an application performs for the end user is a more relevant proof point to be checking. In contrast, at the other end of the scale (in the wholesale world, for example), carriers that are buying from another carrier may not even care about the network SLA as long as the price-per-megabit is attractive and there is ample available bandwidth to throw at any slowdown or outage. For today’s demands on IP telephony, business video communications, unified communications and transactional applications, however, the guarantees need to be on the actual response times and application performance. Thus, now is a good time to think about the SLAs on your data WAN service and whether the service provider is standing by its word on how the necessary applications are running across the WAN, rather than measuring packet throughput and performance, which we would argue is of secondary importance.