Real World Hybrid Operations and Apps on AWS

Editor's Notes

• #2: Questions and Feedback? karimh@amazon.com Duration: ~60 mins.
• #4: This session will provide a 100-200 level introduction into hybrid on AWS. We’ll talk about the AWS hybrid story And we’ll split the body of the session into two key parts Implementing hybrid ops Implementing apps Then we’ll wrap it up …NEXT
• #5: Hybrid Ops provide the foundations, including networking, security and management tools Hybrid Apps sit on top of these foundations. Later in the session we wil talk about the common patterns for hybrid apps
• #6: At AWS is not an all or nothing proposition.
• #8: There are four dimensions to this. Security, is our #1 priority at AWS and we have invested heavily into improving security for all customers Scale, 10x more than the others, Breadth of services, more than 50 at last count Depth of features, we are adding more than 1 new feature every single day.
• #9: Security certifications apply globally to all of our DCs Security is also not a static craft at AWS. We are constantly evolvong and just recently we announce we have achieved ISO27018 certification which relates to data privacy in the cloud We also announced a number of new security products, including a security assessment service Amazon Inpsect, a WAF and a cloud governance solution called AWS config rules
• #11: In 2015 we have already surpassed this number And at our annual confernece “re:invent” just a few weeks ago we launched a number of entirely new products
• #12: Complete technology and enterprise services platform Compute Server Storage Networking Security Enterprise Apps for email, virtual desktops, directory services Mobile development tools, scalable Analytics infrastructure and now reporting tools like Quicksight, Marketplace offers easy software and infrastructure procurement Enterprise Support, professional services and training
• #13: So how do you capitalise on these products and sevrices. In this section we will talk about how companies start by laying the foundations for a hybrid operating model.
• #14: This is the type of plumbing that enterprises have told us they need to implement The first thing that is critical is secure, reliable networking Next, we need to make it easy for companies to connect their access management systems like Active Directory to the cloud Then, enterprises need tools that work with both AWS and their on-premise infrastructure Finally, monitoring tools that can keep your services running smoothly
• #18: As a best practice we don’t disclose locations of our data centers, and so we use peering partners like Equinix who provide demarcation point and cross connect into our cloud services. DX is a Layer 3 service between AWS and your co-lo. BGP and HSRP for failover etc. Customers can use multiple DX and VPNs for redundancy and BGP/HSRP to manage failover
• #19: If you have additional security products in your on-premise envioronment you can bring them along with you. AWS Marketplace makes procuring and implementing these solutions a one-click experience. Licensing is handled by marketplace and fees are automatically charged to your AWS account. Once purchased Marketplace install the solution into the VPC of your choice. There are thousands of products across multiple categories like security, business intelligence and management in AWS Marketplcaae
• #20: So I’ve shown you how easy it is to setup a high performance, secure and scalable connectivity and networking between AWS and your existing environment using DirectConnect and VPC. Let’s move onto the next step …. NEXT
• #21: Secure access management is the next step…NEXT
• #22: AWS Directory Service is a full managed directory service that operates in two modes. One is as a standadlone SAMBA4 compatible directory. The other is as a Connector (or Proxy) to your Active Directory environment. With a few clicks you can have your corporate AD environment federated with AWS to for secure access management and single-signon. This means your AD identities can be associated with permissions to AWS platform services. Such as the ability to create and terminate EC2 instances. In CloudTrail, which keeps a log of activities – down to the API level – you will be able to sit individual AD user and what action they attempted to perform.
• #23: IAM is our policy driven access management system. It is extremely fine grained. You can use IAMs built in identities, or federate, to link IAM to your existing AD environment
• #25: Using AWS Directory Service and AD Connector it takes just 15 minutes to setup!
• #26: For those who want to use their existing Identity Management solutions, we have partnered with all the leading vendors, including those which you can see here
• #29: Almost every company we work with that is implementing a hybrid strategy is actually choosing just one cloud to partner with. Customers tell us they prefer a cloud partner with a large breadth and depth of services. Multi-cloud is being talked about as a potential long term option, but they are all starting with one knowing that the beginning of a long journey starts with a few humble footsteps. Cloud brokers, or systems that abstract customers from cloud providers APIs, products and services, negate just about all of the benefits of the cloud. Broker platforms are slow to respond to or never support new services, features and API. They can become bottlenecks and introduce reliability issues. In fact, many brokers stop at “VM” integration which is simply not enough.
• #30: Open source has well and truly arrived and it is an extremely viable, and in many cases better alternative to commercial solutions. OS is free, often field tested at scale and is updated frequently. Some of the worlds largest startups like Netflix use AWS and are building ecosystems of management tools to add additional features and capabilities.
• #31: Enterprises are using tools like these to fully automate the deployment of their applications HASHICORP Packer: solves the problem of building multi-platform (VMWare, AWS) machine images Terraform: allows you to describe an application stack (e.g. 3 tier) as a template and then use that template to deploy app environments Ansible: a popular configuration manegment tooll for scripting changes to virtual machines, EC2 instances and applications
• #32: Describe the basic machine type (specifications, running services tc. and types of images you want to produce “builders”) and packer automatically generates those images (AMI, VMX etc.)
• #33: The config follows the industry standard JSON format If you were using VMWare you would create another builder config with the desired specification
• #34: Terraform uses a similar JSON based format. This script for an application running on AWS creates 5 amazon EC2 instances of “m1.small” type and puts them behind a AWS manage load balancer on port 80 and routes through to internal port of 8000.
• #35: This is a common workflow in enterprise environments. Let’s walk through it
• #36: If you have existing VM images you can easily migrate them to AWS using these tools
• #37: What if you wanted to move lots of virtual machines or evne an entire data center That’s where Snowball comes in. Keys are not stored on the appliance It can withstand a 6 G jolt
• #38: So we have talked about how packer, terraform, ansible can help you implement automated builds across your on-prem and AWS environments
• #39: Now let’s talk about monitoring tools
• #40: AWS comes with a built-in monitoring platform called CloudWatch. CloudWatch provides useful dashboards and metrics and simple alerts for our platform services and features. Many of these are at the infrastructure level. But you WILL need more detail than that, so you want to invest or textend in your existing tools. New Relic and App dynamics provide application and infrastructure performance management. And data dog provides really detailed infrastructure. These vendors all have integration with AWS and are mutli platform supporting AWS cloud and on-prem infrastructure Splunk App covers cloudtrail, AWS config, EC2 ++
• #41: Now that you are paying as you go, it’s important to track your spend. AWS provides native tools to do so, and its easy to setup billing alerts. Many of our enterprise customers also use a produc called Cloudability” which gives them detailed spending reports, usage analytics and tools to govern and regulate spending
• #42: Finally, we added application performance monitoring and integrated splunk for operational analytics
• #43: And you’re done.
• #45: Enterprises start with dev and test because it’s a low risk and high return prospect
• #46: More performant: Using AWS, we decreased average network latency from 700 milliseconds to less than 50 milliseconds.” Figure 1 demonstrates the ESS typeahead suggestion service running on AWS. They push servers from 30% utilization up to 70% utilization. Scale your testing: Blue green deployments – multiple live production environments
• #47: Create application templates, implement CI, CD and make DevOps a reality. Test full scale environments.
• #48: One thing I hear over and over again from developerss that there simply arnet enough development environments to go around Developers want more, and they want their own discrete environments. And they want environments that replicate the performance dynamics of the production environment. What about a fully built and managed, data center clustered Oracle 12 database in 15 mins? Or SQL, Mysql, Postgres or even MariaDB? And if you’re outsourcing development you also need to provide a secure workspace. With Workspaces, our virrtual desktop as a service, its easy to do this. Plus with the power of VPC, you can provide network isolated environments. J&J is in the process of rollut out 25,000 Amazon Workspaces for temporary employees
• #49: AWS have developer focused services for enterprises and startups. Even if you’re not doing continuous integration and delivery, you can still harvest the benefits of our developer tools. CodeCommit, Git compatible repository fully managed (we handle storage scalability, security and availability) CodeDeploy to managed deployments. This is based on a tool named Apollo that Amazon has been buulding over the last 10 years as part of our journey towards micro-services.
• #50: The next step some customers take is to move Disaster Recovery environments to the cloud. Because enterprises are paying a lot of money for secondary sites that they seldom use. With AWS, you pay for what you use when you’re using it. Not all customers move DR, in fact many skip it and move straight into production workloads.
• #51: https://aws.amazon.com/solutions/case-studies/delaware-north/ x months into its cloud migration, Delaware North was realizing benefits in addition to its data center consolidation, including cost-effective security compliance, enhanced disaster recovery, and faster deployment times for new services. “Robust security in a retail environment is critical for us because of our many retail operations, and AWS is enormously helpful for that,” says Brian Mercer, the senior software architect for the project. “By leveraging the security best practices of AWS, we’ve been able to eliminate a lot of compliance tasks that in the past took up valuable time and money.” He adds that the company also has increased its disaster recovery capabilities at a lower cost than what was available in its previous data center deployment. “It significantly improved our business continuity capabilities, including seamless failovers,” he says. The solution is also helping Delaware North operate with greater speed and agility. For example, it can bring in new businesses—either through contracts or acquisitions—and get them online much faster than in the past by eliminating the need for traditional IT procurement and provisioning. It used to take between two and three weeks to provision new business units; now it takes one day. The Delaware North IT team is also using AWS to overhaul its operations by eliminating outdated and cumbersome processes, cleaning up documentation, and leveraging the benefits of running test and development tasks in combination with rapid deployment of services through the cloud. “Our DevOps team can now spin up the resources to push out a service in just minutes, compared to the weeks it used to take,” says Scott Mercer. “With AWS, we can respond much faster to business needs. And we can start repurposing time and resources to deliver more value and services to our internal teams and to our customers.”
• #53: If you decomission tape altogether, at the very least send your daily, weekly backups to AWS. Companies like CommVault and Symantec have made that easy with NATIVE integration for AWS S3 as a storage provider. Even if you doint use those products you can use the AWS Storage Gateway, a virtual appliance, that asynchronously replicates data to Amazon S3. Likewise, ordinary server volumes mounted across iSCSI can be replicated into the AWS cloud enabling you to protect specific systems or datasets attached to servers (like databases)
• #54: What about live replicating virtual machines and apps to AWS for DR? That’s possible too Racemi and CLoudVelox both provide solutions to do this and to failover in the event of an issue. Or consider this as a simple and cost effective insurance policy. Use VM Export to make a monthly copy of all of your VMs to a Snowball device and then import them back into the cloud as on-standby virtual machine instances.
• #55: So you got through dev and test and DR… and an amazing new digital project has come up. First off, digital projects are the BEST place to start with AWS … so why is that
• #56: AWS is digital. In fact we have a lot of digital services, like IoT, machine learning, data analytics – but we wont have time to cover all of them in this session – so I’ll just focus on a few
• #57: Adobe launched their cloud computing offering on AWS and one of the primary reasons for that is the level of automation the AWS platform provides.
• #58: We have a large variety of services. If you’re building a new mobile app from scratch then check out Mobile Hub. It This is designed to help you ship your product faster by providing you with all the basic mobile app functionality including identity management and session storage, analytics, content delivery, and app testing. If you have an existing app, or a project already underway, you should definitely check out Elastic Beanstalk. With EB, just upload the code and AWS does the rest, buidling the infrastructure (LB app and database) and scaling to adapt to peaks in demand. Apps can take as little as 2 minutes to deploy and it supports all of the major languages and container types.
• #59: If you want to implement an app architecture that uses both on-prem and cloud, then you can – using DirectConnect. Just remember to keep the app and database as together for a better customer experience.
• #60: For streaming analytics we added two products recently – Amazon Kinesis Firehose which make it easier to get data into Redshift and Amazon S3 and Kinesis Analytics which enables analysts to query streaming data using standard SQL queries.
• #62: So today we’ve talked about the 4 steps to enable Hybrid Operations and some of the really common initial use cases for enterprises.
• #63: Finally, lets stay in touch…please connect with me on LinkedIn Thank you