As defined by NIST, cloud computing is “a model for enabling ubiquitous, convenient, on-demand network access to a shared pool of configurable computing resources (e.g., networks, servers, storage, applications and services) that can be rapidly provisioned and released with minimal management effort or service provider interaction identified by 5 essential characteristics - On-demand self-service, Broad network access, Resource pooling, Rapid elasticity, Measured service. NIST further talks about cloud and explain 3 service models viz IaaS, PaaS and SaaS and 4 deployment models viz. Private Cloud, Public Cloud, Hybrid Cloud and Community Cloud.
This definition is very useful in architecting the approach for selecting a cloud provider and also to manage the SLAs as agreed by the cloud providers.
Essentially, we are talking of tracking the SLAs to meet 5 essential characteristics and 3 service models while dealing with a cloud provider.
Strategy: Migrating to Cloud
Cloud is certainly not meant for each and every need and not for each and every business. Further, business needs to understand that private clouds are very expensive to build and there has to be at least 100 valid reasons justifying building a private cloud.
Business should build a strong strategy before moving on to the cloud and should certainly seek professional advice in designing such strategies as it has long lasting impact on business and technologies and involves people from both areas.
e-Zest suggest to consider the following as a part of strategy building process for enterprises:
- Enterprise maturity assessment: According to Gartner, Maturity Assessment for Application Organizations enables CIOs, application leaders and managers to assess the critical disciplines that are key indicators of the overall maturity of the application organization.Gartner suggests enterprises to mature to level 3 before planning to migrate their applications on the cloud.
- Enterprise cloud readiness assessment: Not all applications would be ready to move on the cloud. A careful inspection from various perspective needs to be carried out such as licensing model, multi-tenancy support, security, elasticity etc. e-Zest recommends a methodical assessment of enterprise applications in order to design the cloud migration strategy. Following are the important activities to be performed by cloud experts as a part of this process.
- Review current enterprise applications
- Data collection through tools and interviews
- Analyze apps using multi-dimensional analyses and best practices
- Evaluate cloud readiness of each application in scope
- Recommend adoption strategy for Private/Hybrid/Public clouds
e-Zest has developed its own tool to help cloud experts collect and process vital information for cloud migration.
- Cloud Migration Process: Once the applications are identified as right candidates for migrating on to the cloud, the next step is to start planning, porting and launching the applications.
- Cloud Management Process: The last and the ongoing process in cloud adoption is to maintain, sustain and manage cloud applications with respect to their functionalities, security, availability and performance. This is part of typical “managed service for cloud applications”. This includes Monitoring of the performance, availability, utilization and costing of the cloud environment to determine utilization, SLA compliance, security status, and optimization thresholds.
Best Practices: Cloud Management Process
Cloud management involves managing the stuff at all the 3 layers of cloud – Infrastructure, Platform and Software.
- Management of cloud at Infrastructure Level: Most of the SLAs at infrastructure level are managed and guaranteed by the cloud service provider. Example: Amazon Web Services provides SLAs for all their AWS resources. For EC2 instance, AWS provides guarantee of 99.95% availability during the service year. AWS gives a service credit in case EC2 doesn’t meet this SLA. Similarly AWS provides SLAs for their other resources such as S3, Route 53 etc. Managing SLAs for various cloud resources in absence of a tool may become complex over a period of time.Apart from managing SLAs at infrastructure level, enterprises have to manage the infrastructure itself. The main challenge in managing cloud infrastructure is complex deployment model, inter-zone data communication, troubleshooting the performance bottleneck, evaluating the elasticity requirements and implementing auto scaling. Traditional way of managing IT infrastructure may not be very useful in effectively managing the cloud infrastructure.e-Zest recommends to have a trained systems administration team to take care of your cloud infrastructure. e-Zest provides “managed services” for all your cloud management needs.
- Management of cloud at Platform Level: Many cloud providers provide various platforms as a service to their users. For example, Amazon Web Services provide SWF (Simple Workflow Framework) which is a workflow service for building scalable, resilient applications. This service provides faster and cost effective ways to build and track processing steps that run at different times and have different durations, while ensuring they are executed reliably and without duplication. Many cloud providers DO NOT provide any SLA for platforms. This poses a great challenge in managing functionality, security and availability related to PaaS. Only way to track such parameters is either by using the APIs provided by the cloud providers or to use some highly matured APM tool that can hook with the underlying request/responses between platform and application and platform and infrastructure.
- Management of cloud at Application Level: Since applications are built on top of platform/infrastructure provided by cloud providers, managing applications and SLAs related to application is even more complex than managing those at platform and infrastructure level. The enterprises need to track the following SLAs for their applications besides desired functionality –
a. Application Availability
b. Application Response Time (Performance)
c. Application Security
d. Total Cost of Owning the Application
No cloud provider provides SLA tracking capabilities at application level since it’s an agreement between application provider and the enterprise seeking application development services. The challenge which both the parties face in tracking SLAs at application level is in isolating the root cause of non-availability, or low performance or security because they have to manually analyze the time critical data at all the levels of cloud.
How e-Zest can help you manage your SLAs in AWS cloud?
e-Zest uses Lantana - its own APM solution for AWS. Lantana is a complete application performance management solution for AWS based applications with tons of features. While it provides deep and actionable insights about AWS based applications, customers don't have to install anything. Its capable of performing auto-discovery of the AWS resources, generating alerts and notification and provides a monitoring solution for the application support team.
Lantana is differentiated with the following key features:
- Frustration Free Start: Lantana is completely agentless, SaaS based offerings. It doesn’t need anything complex to use it to monitor the AWS based applications. It’s intelligent and sets up itself using auto-discovery and auto-configuration features. It needs just AWS credentials to start and system is ready for you to manage your applications on AWS.
- Self-Benchmarking: Lantana provides a very innovative feature of self-benchmarking. Performance charts to benchmark against your application’s past performance gives you real time indicator as to how your applications should perform in various conditions. All the configuration items are monitored in near real time to diagnose problems before impacting users.
- Support Team Friendly: Lantana powers the support team by increasing the efficiency of since they can get the exact cloud performance snapshot at a particular point in time. Support team can get useful information from Lantana in case it finds any performance issues in the AWS based applications. This enables support team to quickly fix application performance issues proactively without letting it down due to over utilization.
- We understand AWS: AWS is evolving as PaaS. Its lot more than on demand EC2 scaling. This is where we have designed the whole product around AWS to provide the holistic view of AWS assets as against fragmented picture. Our team comes with the past enriching experience of SOA, other clouds and Amazon.
- Performance Analytics: Get various dashboards from historical data which you won’t get otherwise from AWS administration console to know the root cause of performance bottlenecks, resource over utilization, and monthly spend. This becomes very handy to make important design and deployment related decisions for architects, business owners and system administrators.
For more information about e-Zest, please visit www.e-zest.net
For more information about Lantana, please visit www.lantanacloud.com