INAP IS NOW HORIZONIQ.
Click here to LEARN more.

Jan 22, 2013

Will Cloud SLAs keep pace with the demands of enterprise SaaS?

INAP

Cloud SLAs keep pace with enterprise SaaSBy Charlie Alsmiller, CEO, Appterra

I recently spoke with an enterprise software company that had launched a “Cloud” initiative. Like so many others, ourselves included, they had flocked to the variety of services available from Amazon, Azure, Softlayer, Rackspace and others to “scale” their business using the latest cloud technology.

I have examined all these services in some level of detail, and have found them all to be very good when paired with the right job. In our business, we rely on our technology partnerships to help ensure 100% uptime for our customers. In turn, our customers rely on us for dependable service that allows their business to run smoothly. Downtime is a really big problem. As such, they expect us to write a fairly intense SLA (Service Level Agreement) into our contracts, because if we are down, they are down. There are usually substantial financial implications of downtime, and it’s important to ensure that all parties are fully “incented” to react quickly if a downtime situation arises.

As we all know, downtime has been a problem with many of the larger providers recently. I understand that downtime happens, but where this gets to be a challenge is in the SLAs of these firms. Are you willing to bet your business on this? Below is an excerpt from a leading firm’s Service Level Agreement, which is typical.

“XXXXXX will use commercially reasonable efforts to make XXXXXX available with an Annual Uptime Percentage (defined below) of at least 99.95% during the Service Year. In the event XXXXXX does not meet the Annual Uptime Percentage commitment, you will be eligible to receive a Service Credit.”

This firm goes on to note that the service credit will equal 10% of your bill for the affected period. Thanks.

I have a former customer that once calculated downtime costs for his business at $1000/minute. I also have a current customer that received EU 500/minute penalties for downtime in their ability to communicate and take orders from their customers. So, let’s run some quick math. 99.5% uptime translates into approximately four hours and 23 minutes of downtime per year, or around 22 minutes per month. In the case of my former customer, 22 minutes of downtime per month equals $22,000 of lost revenue. If a business pays $10,000/month for cloud services, they would only receive a $1000 credit for their loss. That’s not acceptable, and I can assure you, they won’t wait around for it to happen again – they WILL look elsewhere.

The bottom line is that while the outage will have financial impact, both in terms of lost revenue and SLA payouts, the major hit will be that of customer confidence. Any credits offered will be insignificant in replacing either revenue or customer perception.

As a SaaS company, we have a lot of moving parts — our network, hardware, operating systems, database systems, our code base, etc. All must work flawlessly together to ensure a great end customer experience. As enterprises seriously consider cloud services, take a cold, hard look at the Service Level Agreements and the real commitment to uptime that the supplier can provide. I think you will be surprised and choose your partner very carefully — which is why we chose Internap as our cloud hosting provider.

For more information about evaluating cloud providers, read the Cloud Hosting Buyer’s Guide.

Explore HorizonIQ
Bare Metal

LEARN MORE

About Author

INAP

Read More
Jun 1, 2012

No compromises: performance and reliability

INAP

No compromises: performance and reliabilityA couple of my previous blogs have focused on two key aspects of the sport of Formula 1 racing: testing and teamwork. Both of these points lend themselves to better performance, and F1 is all about performance — but just as important is reliability.

Specific manufacturing methods can give an F1 team parts that will have the best performance available but can only be used one or two times. This means the part is high on performance but low on reliability. On the other hand, F1 teams can also have very reliable parts manufactured, but not gain all the performance from the part needed to win races. This results in high reliability and low performance.

F1 teams need to balance the need for performance (optimization) with the required reliability (durability), based on the FIA rulebook. Here are a couple examples from the FIA rulebook:

Engine limitations: The FIA rulebook states that teams are allowed eight engines total during the season, and that each engine must be used a minimum of two races. If the team tries to use more than eight engines during the season, or switches out an engine after one Grand Prix weekend, the team is penalized. The penalty is usually in the form of the car being pushed down 10 places on the starting grid.

Transmission limitations: The teams are also required to use the same transmission/gearbox for four races. If the team trades out a transmission before the allotted number of races, they are penalized. The penalty is usually in the form of that car being pushed down the starting grid a certain number of positions, usually five.

In each of these scenarios it is all about being able to master this delicate balance of getting the most performance available out of each part and have the part last to gain an edge over other teams. The job of the IT professional is not unlike that of an F1 team – it’s a constant challenge to maintain the right levels of performance while still ensuring your infrastructure is reliably supporting your business. And there really isn’t much room for compromise when it comes to your critical infrastructure. When looking for an IT provider to help you meet those challenges, it’s important to keep a few questions in mind:

  1. What level of availability do they offer? Meaning what is their uptime percentage?
  2. What sort of uptime guarantees do they offer in their Service Level Agreement (SLA)?
  3. Do they offer proactive credits in the SLA? Or do they require you contact them first?

These prospective provider questions and more can be found in many of our buyer’s guide eBooks. Whether you are looking for advice on cloud, colocation, hosting, content delivery or accelerated IP, we’ve got you covered. Just scroll down to the eBooks section on our resources page for some useful summer reading.

Explore HorizonIQ
Bare Metal

LEARN MORE

About Author

INAP

Read More