The founders of ElasticFusion have helped some of the world’s largest corporations with their cloud deployment strategies and implementations. Along the way we have learned what works, what doesn't work and what new technologies present the promise of greater agility and availability and merit investment to enable businesses. All done with security in mind from the start.
We work with your staff to transform them into an agile, public cloud adept team that is competent with the tools chosen for the project and poised to successfully tackle the next generation projects.
We focus on patterns that are repeatable, verifiable and secure. Operationally, our number one priority is the preservation of data. We build everything else around that basic consideration.
Compute models will vary by need or SLA but all are built to minimize cost and maximize availability. Our goal is to work with our clients to build a framework that supports a well monitored, self-healing, auto-scaling and as cost-efficient compute models of delivering application value as the constraints the client sets allow.
ElasticFusion begins with security. We work with our clients to integrate security into their cloud implementation from the start.
Security is a consideration, while we work in an agile framework, before the first line of code is written. For projects already underway, or for legacy code uplifts, a comprehensive review is done to ensure that it can be safely implemented in a public environment.
These reviews are ongoing. Public cloud best practices are fully incorporated and when the project is sufficiently mature, penetration testing is performed.
Compliance is also integrated into the process, whether your requirements, PCI DDS, HIPPA, Dodd-Frank, SOX, ITAR we have experience with locking with controlling the public cloud environment and providing the reporting apparatus required to satisfy the regulators and compliance initiatives within your organization.
It's a rare thing when we get to start a project in the public cloud with a greenfield project. A project specifically cloud built. It happens and it should be savored.
More often we are asked to uplift an application built for a terrestrial data center and terrestrial data center presumptions. Always on. Limited horizontal scalability. A relational database or a file system are the only means of storage and so limited vertical scalability.
The uplift, for a heavily used application, is only the first step. We must reevaluate the architecture and optimize it for the cloud. This may mean changing both the compute and or the storage models. For example, this may mean re-architecting it as a microservices built using serverless philosophy or Kubernetes depending upon business goals.
Other applications, a low use application, should also be reexamined. The goal here is to minimize the changes. But we may need to ask hard questions about true availability requirements. For example, would a 98% SLA be acceptable?
As we triage all of the uplifts, we will drive down the costs to an acceptable level, in a way that a straight uplift and stop will always fail to do.
Many organizations have solved the problem of continuous integration. Don't worry if your organization has not, we can put that together for you. Continuous deployment is harder remains a goal for many.
Our team will guide your team to complete CI/CD deployments using standard tools.
Open Source or public cloud tools in our kit include: Jenkins for software builds. Maven and Gradle for Java specific implementations. Chef, Puppet or Ansible for server builds. Ansible, TerraForm or CloudFormation for Infrastructure as Code. Docker, Windows Server Containers and AWS ECS or Kubernetes for deploying the containers. Monitoring with Sensu, CloudWatch, Monitor and/or Stackdriver. OSSEC for Security Monitoring. ELK for analytics and monitoring. Vault and credtash for secrets management.
RunBooks are workaround.
They are a reflection of your DevOps backlog.
See a problem. Solve a problem. Document the solution. Reflect the solution in code.
Sounds straightforward. Of course, it isn't. But the sooner your organization prioritizes these sorts of backlog items, the lower the total cost of the monitoring team.
