OpenCredo Cloud Report: July 2017

Up to this point, the cloud services market has been heavily biased towards migration from on-premise to Infrastructure-as-a-Service (IaaS): comprising compute, networking and storage resources which:

are available in abundance
can be provisioned and de-provisioned on-demand
are billed by utilisation
are programmable and can be provisioned with code
can be provisioned globally, in isolated regions and zones
have security baked into the physical, networking and virtualisation layers

This core vision has driven the first wave of cloud adoption as organisations look to gain on-demand scalability, alignment of cost with utilisation, increased agility and migration from CapEx to OpEx. OpenCredo have successfully assisted organisations big and small in migrating their workloads and building applications which align natively with the cloud – as well as automating their infrastructure and delivery pipelines. The OpenCredo cloud practice sets out to cement our existing work with this core cloud vision, whilst helping our clients realise opportunities presented by recent developments in the public cloud.

Recent Developments

Amazon Web Services (AWS), Microsoft Azure and Google Cloud Platform (GCP) have emerged as the 3 major players in the public cloud market. These organisations are leveraging their already significant investment in infrastructure to create public cloud platforms with features that are almost impossible for on-premise systems or smaller cloud players to match. From private intercontinental cabling and bespoke hardware through to elastic distributed systems provided as a service, the cloud vendors are rapidly innovating and differentiating from each other.

Some of the emerging trends are:

Container Standardisation

Containers have emerged as key vehicles for the seamless movement of workloads across multi-cloud and on-premise environments. With increasing consensus around the Docker container format and Kubernetes orchestration platform and early moves towards federation, we are now seeing sufficient standardisation for workloads to be successfully deployed across environments. OpenCredo have been working with containers for several years – including sponsorship of the ContainerSched conference and partnerships with Google (Kubernetes), RedHat (OpenShift) and Mesosphere (Mesos DC/OS).

Serverless Function-as-a-Service (FaaS)

Lead by AWS Lambda the FaaS paradigm delivers systems which execute code (functions) in response to internal and external events. These functions are uploaded to the cloud system and executed in a fully abstracted environment – there are no servers to provision or scaling concerns. FaaS is driving novel application architectures, infrastructure & security governance and data systems.

Serverless Services

Other “serverless” services are emerging. These are generally characterised by being auto-scaling and self-managing. Many of these are in the data space such as BigQuery (GCP) and Athena (AWS) and allow data analysts to focus on their tasks without having to worry about the underlying infrastructure.

Observability

With serverless, containers and the increasingly use of distributed architectures such as microservices, it becomes ever more difficult to understand the current status of the overall system. We are seeing innovation in the monitoring space with distributed tracing tools such as Zipkin and AWS X-Ray and container friendly monitoring such as Prometheus and SysDig. Even with advanced tooling, it remains a strong UX challenge to understand how different personas need to interact with the cloud system.

Security and Compliance

All three cloud vendors are growing their service offerings around security and compliance. In some cases, this means more transparency around compliance certification and in others additional services for key management, encryption and logging. In many cases, the security services natively on offer allow controls which far exceed those available on premise. AWS Lambda and Config in particular allow business rules, logic and code to be executed on every infrastructure event. Combined with a growing ecology of third party vendors, security for even highly sensitive data becomes possible.

Machine Learning

Machine learning is undergoing a renaissance and the cloud vendors are very much onboard. All three provide APIs which allow access to pretrained models – for example the Google Vision API classifies images into categories, detects objects in the images and reads text. In addition, each platform has its own managed machine learning framework – such as TensorFlow (GCP) and MXNet (AWS).

Specialist Hardware

Machine learning and high performance computing is highly compute intensive. We are seeing a range of hardware innovations in the cloud which forge a new direction away from the previous proposition of horizontally scalable commodity compute. These include GPU, TPU and FPGA.

As opaque managed services, many of these developments challenge OpenCredo’s heritage as supporter of Open Source. In our opinion, however, the business benefits of these services are undeniable – in some cases shifting the burden of operations to the cloud and in others accessing innovative approaches to delivery. We are heavily investing in cloud services to help our clients understand the nature of the services, lock-in and comparable Open Source alternatives. With auto-scaling, TCO and cost management also become essential.

Why OpenCredo?

OpenCredo has proven expertise in building systems which exploit the benefits of the Cloud. By developing strong partnerships with leading technology providers whilst remaining solidly independent, we are able to present a balanced view about your adoption strategy. All our consultants are also hands-on technologists so we are able to provide you with a no-nonsense view of technologies from genuine practical experience.

We aim to help you make fundamental business decisions on: