The SLAs don’t work: How to drive better performance in the cloud

With more and more mission-critical workloads now moving to the cloud, it is vital that companies have confidence in the ability of their cloud service to support the needs of their business, whilst also providing value for money. To put it simply, would you rent a car before checking your insurance liabilities and contract? Probably not, yet many businesses are often prepared to launch a cloud service with little assurance of its capabilities. The harsh reality is that, whether you’re an enterprise creating your own private cloud or using a service provider delivering public cloud services, poor performance will immediately erode customer trust; reduce end user adoption; and ultimately impact company revenue.

Hitting a bump in the road

It is an unavoidable truth that infallibility is impossible when it comes to the cloud. While a cloud service may promise high availability – for example, Microsoft guarantees 99.99% availability – this still leaves a lot of margin for downtime that businesses cannot afford. Major outages, such as that recently suffered by Microsoft when its Azure cloud service went down, prove that no provider is immune to experiencing an outage; which can potentially cost around £3,000 a minute. What’s more, this type of outage is not an irregular occurrence. According to Outage Analyser, which tracks internet service outages, Azure has regularly suffered outages, which impacts up to 1,234 websites each time – causing headaches for other businesses reliant on its service.

However, it’s not just the availability of a cloud service that is important; it’s also its performance, which can impact the experience for end-users. While a cloud service may well be ‘up and running’, it may also be performing at a painfully slow speed. Unfortunately, the majority of cloud providers only offer an SLA that guarantees customers certain levels of availability. Measures around performance are less common, leaving customers to take a leap of faith and forced into a ‘take my word for it’ scenario – which can ultimately be detrimental to a business’s operations, should performance problems occur.

Driving without insurance

Perhaps tellingly, a recent survey reveals that the majority of businesses believe the SLAs offered by cloud service providers are too basic and fail to address the risks they face as a result of moving to the cloud. Research also shows that 60% of businesses using cloud services worry that co-located tenants could negatively impact the performance of their applications by monopolising shared resources. This ‘noisy neighbour’ effect can be a particular problem during peak hours of the day, such as when businesses are coming online in the morning and employees are all logging on at the same time. In much the same way that hotel managers need to ensure there’s enough hot water for everyone regardless of how many are taking a shower at the same time, cloud providers need the ability to guarantee the consistency of computing resources and ensure the demands of one don’t impact upon the many.

Get ready to race

With the lack of confidence in the guarantees offered by cloud providers apparent, customers must demand granular SLA assurances around performance and rapid problem resolution. To get on the right track, there are three basic measures that businesses should take to ensure greater stability behind applications and workloads in the cloud:

1. Businesses should insist on SLAs that are measurable against their objectives, not the cloud providers’ reporting needs. By defining expectations in the context of what is important to the business, companies can set the parameters of their SLAs to provide the greatest protection to the most important areas of the business. They can also determine how they will be notified and what the committed response times and escalation paths are in the event of a service problem.

2. In an age when customer loyalty is fleeting, a poor end-user experience can quickly equate to lost revenue. Every user and every transaction is essential, so a monthly availability report fails to provide the insights needed to protect revenues and react quickly with proactive support for those experiencing problems. Businesses need 24/7 visibility into performance from the perspective of every single user to ensure that nobody falls through the cracks.

3. Businesses should have visibility into public cloud infrastructure, with the ability to identify network bottlenecks and track resource utilisation. This insight will empower them to challenge their cloud provider over resource availability and over-utilisation that could be impacting on the quality of experience their end-users are receiving. Businesses should also tailor their SLAs to put the onus on cloud providers to identify and move noisy neighbours before they impact performance.

Crossing the finish line

It is clear that customers are calling for cloud providers to guarantee more than just availability. The performance of a cloud service is equally as important, and can massively impact the operations of a business. With this in mind, the everyday cloud provider has a long way to go before their SLAs are aligned with the needs of customers. In the meantime, businesses can empower themselves to hold their provider to account by monitoring the performance of their cloud services proactively, rather than relying on the often obscured reports offered by providers. This will put businesses back in the driving seat, and in a better position to realise the true potential of the cloud.



Type: White Paper


  • Favorite list is empty.
FavoriteLoadingClear favorites

Your favorite posts saved to your browsers cookies. If you clear cookies also favorite posts will be deleted.