Why multi-tenancy architecture is the new normal
Sumo Logic CEO Ramin Sayar delivered his keynote at the company’s Illuminate 2019 conference to reinforce his assertion that there is ‘a massive intelligence gap’ throughout modern technology stacks caused by the pressure that comes from ‘continuous release’.
The world of always-on web-connected cloud-services driving mobile-optimised applications is one where users want applications to be continuously available, continuously connected, continuously updated, continuously integrated and continuously secured… and so, continuously released.
This drive for continuous release depends upon the existence of cloud computing.
Consequently, is has implications on computing architectures, which must now be built for a world that is inherently multi-tenant on the public cloud.
CEO Sayar says that as we move to continuous data, we need to provide users with what he could also call continuous choice. Users want choice over what device, what platform and what application they use and so on… plus they also want choice over where their data resides.
Noisy complexity
“But choice creates ‘noise’ and complexity. Most enterprises today are using up to somewhere around 30 tools to perform monitoring… and the problem is only made worse when we find out that the monitoring tools themselves don’t interconnect,” said Sayar.
The antidote to this predicament is what Sumo Logic likes to call continuous intelligence… and it’s a platform that is true multi-tenant SaaS in nature.
What this leads us to is the suggestion that multi-tenant cloud architectures are essentially the new normal.
“Multi-cloud [built on inherently multi-tenant instances] and open source technologies, specifically Kubernetes, are hand-in-hand dramatically reshaping the future of the modern application stack,” said Kalyan Ramanathan, vice president of product marketing for Sumo Logic. “For companies, the increased adoption of services to enable and secure a multi-cloud strategy are adding more complexity and noise, which current legacy analytics solutions can’t handle. To address this complexity, companies will need a continuous intelligence strategy that consolidates all of their data into a single pane of glass to close the intelligence gap,” he added.
But for this level of intelligence to have substance, we need to think about how it is applied at the coalface of the modern IT stack.
Continuous Intelligence (Cont-Intel)
Ramanathan says that for Continuous Integration (CI) and Continuous Delivery (CD) to run effectively, it needs Continuous Intelligence (let’s shorten to Cont-Intel) throughout the entire Software Application Development Lifecycle (SDLC).
He also notes that his firm is ‘fundamentally different’ to other machine-generated log analytics players (Splunk being the obvious suspect) because Sumo Logic was born in and of the cloud (Splunk dates back to 2003 – Sumo Logic is a newer cloud era baby born in 2010)… so, he claims, the company is natively attuned to the service-centric consumption model of the cloud.
Why multi-tenancy matters
When we first started talking about multi-tenancy in cloud computing, people were concerned. Customers thought that putting all their data on a shared multi-tenant cloud datacenter server next to somebody else’s data to be pretty irregular.
Today that view has changed.
Today the industry understands that multi-tenant architectures proliferate, typify and exemplify the nature of the modern web with its continuous ubiquity.
But there’s another lesson here… and it’s one for DevSecOps (i.e. DevOps delivered from a core security stance).
“Looking at the state of the market as we have detailed in our Continuous Intelligence report, I would emphasise to you that we say that Sumo Logic is natively multi-tenant… and here’s why multi-tenancy matters, especially to DevOps. When a Denial Of Service attack happens in a company, it doesn’t happen in a pocket, it happens in many places throughout the stack. So being able to correlate issues right throughout the stack becomes fundamentally important,” said Ramanathan.
Ramanathan urges us to think about DevSecOps in slightly different terms. While he’s okay with the term itself, he suggests that it might perhaps more accurately have been labelled DevOpsSec… or perhaps even DevOps_Sec.
The difference is, this would be DevOps that is essentially concerned with operations, but has still has a sharp eye focused on security issues when they occur.
Times have changed, cloud perceptions have moved on, our notion of how we apply analytics to an increasingly complex cloud must (arguably) also move on to. This is where Sumo Logic is at and it’s a subject that is (again, arguably) spiraling upwards in the modern IT stack of today’s now ever-so continuous company.