Guarantee the application SLA 24/7, a complex and costly challenge for a SaaS publisher.

The application SLA, now required 24/7, imposes to publishers a complex and expensive system to set up internally.

The SLA (Service Level Agreement) contractually defines the minimum quality level of a service provided by a service provider to its customer.

In the case of a software publisher, which offers its application in SaaS mode, the application SLA determines the availability of the application online, or its rates of accessibility by users. In other words, it is the maximum time of disruption of the application service tolerated by the client of the software publisher, under penalty of financial penalties.

For example, a 99,5% application SLA will guarantee customers of a SaaS publisher a maximum of 4 hours monthly downtime of its online application. Until now, nothing new... To benefit of the online application in SaaS mode, the customer takes an "all-inclusive" subscription directly from the publisher, which includes the software license, hosting and associated services. It is up to the publisher, the only client's interlocutor, who has contractual and operational responsibility for the SLA application.

“ In the case of an application in SaaS mode, it is the publisher who is responsible for the application SLA...”

A new kind of responsibility for the publisher. The publisher must therefore orchestrate each technical link in the production chain, from the network to the hosting, including security, physical management of servers, systems administration, management of service contracts and maintenance, to be able to consolidate and contractually commit to the resulting indicator: the application SLA.

But in its course to manage its SLA application, the publisher is faced to several successive challenges.

Challenge n°1: ensure an SLA with 24/7 coverage, user standard of use And yes, uses have evolved, increasing the level of customer requirements: organization over several times zones, teleworking, continuous use via mobile, work in staggered hours...

Every company application, whether it's critical for business or not, must now be continuously available.

"24/7" has therefore become the standard for use and also for application SLA!

Organized until then to work over a period of office hours by providing functional support for its software, any publisher must now review the organization of its technical team.

Challenge n°2: stabilize its technical organization, constrained to 24/7 on-call duty

First, and in order to respond quickly to the previous challenge, all or part of the technical team, generally made up of developers, is called to ensure the 24/7 coverage of the hosted application.

Concretely, this involves setting up a first supervision system. The objectif is to notify malfunction alerts. During non-working hours, these alerts are taken care of by the engineers appointed on-call, who sometimes have to wake up in the middle of the night to intervene for several hours.

But this organization is only possible if the activity of the SaaS platform is measured and generates few alerts... Few night interventions are easily accepted by on-call teams thanks to intervention bonus.

With the activity growth of the SaaS platform, the number of alerts increases mechanically, but also the criticality level of the platform, because it supports more and more turnover, increasing pressure on the teams.

“ Preserving its technical team is a major issue for any publisher: any new 24/7 constraint, often repetitive with a little added value, represents a major HR risk for the team stability. ”

Indeed, the on-call duty limits of the technical team is quickly reached: the time devoted to on-call interventions quickly takes precedence over time devoted to the development of new functionalities, the added value of a SaaS editor.

In addition, this leads to wear and demotivation of technical teams, tired of night interventions which most of the time are basic and repetitive technical tasks (reboot of servers to restart a faulty service, etc), without even talk about the productivity drop the day after the intervention.

Challenge n°3: set up a structure dedicated to 24/7 while remaining competitive

With the load evolution on its SaaS platform, the publisher must therefore abandon the fragile model based on on-call penalties to further professionalize its production.

This professionalization, or this industrialization, requires at least a complete incident management solution, which integrates the supervision tools (detection software, real-time alert of technical anomalies) and also above all, a human organization able to intervene in case of incident, on 24/7 coverage, with a high reactivity and good procedures.

“ Providing a dedicated human structure to ensure the 24/7 application SLA quickly becomes inevitable.”

If he decides to set up a such internally solution, the publisher must operates his own application 24/7, which implies:

• the setting up a team of at least 6 people to ensure the 24/7 activity ;

• an advanced experience in supervision ;

• technical skills covering the entire supervised perimeter, up to the expertise level ;

• a work methodology based on a set of flawless procedures ;

• supervision tools and software.

All this represents a significant cost... hardly reputable on customers because the 24/7 application availability is supposed to be already included in the costs.

Not considering the many problems to stabilize the cost of a 24/7 team: turnover and loyalty problems, etc.

Sustaining your own 24/7 team while remaining competitive quickly becomes an impossible mission for a SaaS editor.

Another option is to rely on a specialized service provider offering a turnkey solution in the form of a flat-rate subscription.