Altisource PaaS for Verizon Terremark Cloud
Altisource SaaS applications
Microservices (Foundation)
Cloud IaaS Provider - Verizon, AWS
Document Management, Analytics, Marketplace...
Identity Management, Authorization, Rules Management, Search and Audit, Workflow
Compute, Storage, Public IPs, Perimeter and internal firewall...
We needed a PaaS because...
Every app or service needs at least the following
- Security
- Deployment Automation
- HA across availability zones
- Performance management and Scalability
Security
The number one reason why apps are choosing a private cloud is security. We are addressing this at various levels in the PaaS
Central privileged account management for ssh, vpn and database users
Data-at-rest encryption for all our data stores - MySQL, MongoDB, OpenDJ, Elasticsearch with central key management
Central log management for syslog, platform logs and application logs to provide complete auditability
Intrusion Prevention / Detection
IP White / Blacklisting
Deployment Automation
Applications and Microservices need the ability to deploy to the cloud seamlessly from developer desktop through QA and UAT environments to production
Developer
DevOps
Operations
Vagrant plugin for Verizon
Application Deployment Portal
App Blueprint specifying auto-scale groups and connectivity
Each auto-scale group defines the template VM or Docker container
- Catalog
- Environment and Network details
- Storage requirements (LVM based)
- Infinicenter Layout and Tag info
- Hostname, Static IP, NTP, IPA - base config
- Chef runlist
Artifact Repository
Docker repository
Chef Server
Terremark API
Deployment Validation Framework
Blueprint
HA across availability zones
All persistence platforms with geographically distributed clustering and Global Traffic Manager for failover and failback
Persistence data platforms - MySQL, MongoDB, Redis, OpenDJ, Elasticsearch
Geographically distributed clusters - multiple availability zones
Backup and Recovery of data in replicated NFS stores
Data and logs encrypted at rest
DB user credentials centrally managed
.
.
.
.
Per application deployment
.
.
.
.
.
.
.
.
.
.
.
.
DB as a service - multi-tenant and massively scaled
Two deployment options:
Performance Management and Scalability
...including things like monitoring, auto-scale, central log management, dashboards
- Monitoring platform has auto-scale conditions set on some predicate based on resource, platform or application metric
- Monitoring platform detects threshold and signals Application Deployment Platform to deploy additional capacity
- New VM / Container is provisioned using Terremark API and Chef
- New VM registers itself to Service Registry
- Load balancer is constantly watching Service registry for any changes. When it sees a new node, it reconfigures itself
Altisource Autoscalability Framework (Tanto)
Central Log Management
Syslog
Platform / Runtime Logs
Application logs encrypted by tenant
- Alerts
- Analytics
- Audit
- Trends
Firewall Logs
Log agents available for following runtimes:
- Apache - MySQL
- Tomcat - MongoDB
- Shibboleth - Open DJ
- RabbitMQ - Elasticsearch
Altisource PaaS for Verizon Terremark Cloud
By Kingshuk Dasgupta
