SDN - OpenNetSummit 2013

1, Value proposition
2, real-world use case
3,Market Segments
Data Center - Scalability Data Center Networks
Carriar - Service Provider Data Networks
Enterprise 
- Streamlined and Dynamic, but Not a technology Nor a market
- Standardized platform for Development of Novel Services

why a new net architecture is needed:
not cost-effective;
not agile enough: time to market; complex/manually intensive provisioning/configuration/network management;
not oriented towards services/end users: static configuration/static traffic patterns and policy
hardware/infrastructure oriented: ports/bandwidth/ids/vlan tags etc

SDN architecture: 25'30"-27'50"
Programmability/contralized Intelligence/Abstraction
OpenFlow: 28'10"-30'20"
southbound protocol
break the layer  boundaries(of OSI 7 layer);

value proposition: 33'00"
automation -> OpEx
optimization -> CapEx ->  more intelligent processing and a better utilization of the capacity;(more efficient bandwidth - mutli-tenants, multi networks run on the same infrastructure)
Monetization -> New Revenues (Dynamics, intelligence pricing)

applications of Segments: 35'40"
[Data Center]
1, Virtualization: VM mobility/Domain isolation 
2, Hybrid Clouds: span public and private DC
3, Multi-Tenancy: multiple customers share services
4, Disaster Recovery: Improved visibility and managememt
[Carrier network] more efficient usage of carrier bandwidth, especially mobile/wireless bandwidth;
1, Mobile Edge: seamless roaming for 3G/4G/Wifi
2, service migration to cloud: span public and private cloud
3, service monitoring: automated provisioning / centralized network view
4, traffic steering: content delivery services
[NFV] carrier lead ETSI



<Network Virtualization>

<Carrier> manage fabric instead fo manage boxes
1, Optimization
- network utilization: Performance-on-Demand (PoD) - e.g. NaaS / Efficient multi-layer traffic engineering/fixed access utilization
- Service performamce: netrwork aware load balancing/Mobile-WiFi offload(or n-casting)/traffic steering for Mobile QoE
- operation: largely centralized upgrades, features,services additions/high-fidelity simulation, straightforward push to production;
2, Monetization
- intellegence : better matching of network supply & user/application demand -- maximized economic utility : PoD- e.g. NaaS; QoS

- Colt: L3 CPE virtualization/ NFV

<Data Center>

<Enterprise>
challenge: multi-VLAN / VM mobility (IP addr mgmt, access control)
requirements: Live migration / Traffic prioritization / Policy enforcement/ Ease of operations and maintenance
SDN solution: (NEC+Hyper-V)
• Live VM migration without network reconfiguration
• Integrated virtual and physical network management
• Greater flexibility/faster deployment of network services

• Open Standard – more infrastructure options


<SP-SDN>
1, NTT: traffic on demand(by user) / VLAN-MPLS(Cloud)

2, Ericsson+Telstra: virtualization of the aggregation netowkr (NFV), 
                            IP edge of the network & service chaining(1st SDN use case for SP)
   where treat different data flows different ways & not route traffic in and out of boxes that don't need to.
   so that Operator can 
   A) create service that use less network resources.(lower capex/opex)
   B) Create service more quickly because we can route services more dynamically and quick TTM without traditional integrating to every network and every network box. 
   Using traffic steering as part of SDN allows to individualize products more quickly and more readily.

   Telstra expects to move to field trials later in 2013 and to take the first steps toward operationalizing SDN in its production network in 2014.

   Ericsson tech talk about Sp-SDn: cross domain (Radio/Core/DC)

blog to read:  SP SDN gets real

Difference between SDN & NFV: SDNCentral article

-  http://www.nojitter.com/post/240158913/whats-the-role-of-sdn-and-uc

IaaS/PaaS players to watch 2013

http://shakayumi.wordpress.com/2013/03/04/thirteen-aws-cloud-alternatives-to-watch-in-2/
[IaaS]
- AWS

- Verizon Terremark : more secured solution

Verizon/Terremark has already launched cloud services compliant with with 
   the Health Insurance Portability and Accountability Act (HIPAA), 
   the Federal Information Security Management Act (FISMA), and 
   the National Institutes of Standards and Technology (NIST) 800-53 Security and Privacy Controls. 
Their compliance cloud services include Collocation, Managed Hosting, Enterprise Cloud, Enterprise Cloud Express Edition and Enterprise Cloud Private Edition, for each which Verizon offers a business associate agreement (BAA), which guarantees that the data in question is secure, supports geolocation requirements, and is auditable by tenants.

- CenturyLink/Savvis : security and reliability
Enterprises with mission-critical workloads and/or sensitive data like the idea of cloud for its flexibility, cost savings and automation, but they remain wary of potential security breaches or data loss due to outages. Many of these enterprises shun Amazon Web Services and look to Infrastructure as a Service providers with a history in uptime and reliability, such as CenturyLink/Savvis.
Savvis is also addressing interoperability concerns while making a push for hybrid cloud — having the ability to work with private cloud and move workloads to public cloud when necessary.

- Goolge Compute Engine : the greatest threat to Amazon's dominance in IaaS;
Google’s Compute Engine (GCE) remains in limited preview, and it doesn’t yet support Windows, What makes Google even more dangerous is that it owns fiber-optic networks, unlike Amazon, which rely on ISPs. Google has been ahead of the software-defined networking (SDN) game as well, and can compete with Amazon on pricing and performance right out of the gate.
Amazon is still ahead of Google when reaching out to enterprise customers, but Google could change this dynamic if it chose to. With GCE expected to open up to general availability this year, Google could be a major disruptor in the Infrastructure as a Service market.

- Rackspace: fanatical customer support
Rackspace adjusted its managed cloud support strategy from a pooled model of resources to a more traditional dedicated support team model.

In addition to its managed services work, Rackspace cloud also appears to be inching toward the Platform as a Service (PaaS) realm with its Service Registry, a tool that enables cloud consumers to orchestrate the process of assigning a workload and having a cloud service respond to that workload automatically. 


Rackspace has also played a significant role in shaping OpenStack. The company began offering public cloud services based on OpenStack in August 2012, setting the stage for other major cloud service providers to follow. If Rackspace continues to broaden to its IaaS managed cloud services offerings as well as evolve its PaaS products for the burgeoning DevOps market, it should have an exciting year. 

- SoftLayer : Bare Metal Cloud 
The Dallas-based company, one of the largest privately owned cloud infrastructure providers, believes its competitive advantage lies in its array of product options and automation offerings designed to more finely customize cloud environments.

One of those product options is the “bare metal cloud,” in which SoftLayer removes the hypervisor from the mix and offers customers a choice in customizing its hardware infrastructure, including one to 64 processors and access to solid-state drive (SSD) storage and a high-speed global network that can be provisioned in real time. SoftLayer’s technology doesn’t come with lengthy commitments, which minimizes users’ financial risk.

A lot of users like the bare metal cloud, where SoftLayer just give them the raw power for running ‘big data’ or large databases. SoftLayer focus on customers that have performance issues where they need high disk I/O and higher network speeds and AWS can't meet this need.


Its CloudLayer services can either be integrated with dedicated servers and services; they also operate in a standalone mode that can be delivered on demand and self-managed. It isn't as big as AWS from a scale perspective, but they have more options available to users that let them fine-tune their cloud.

- ProfitBricks : Speed/InfiniBand
By adopting the venerable InfiniBand protocol, a switched fabric communications link more traditionally used in high-performance computing (HPC), company officials believe ProfitBricks has a significant speed and performance advantage over its competitors.

We were the first to bring InfiniBand to IaaS market, which changes the dynamics in a couple of ways: through higher performance and greater vertical scalability — the whole up and down provisioning aspect; and we are better able to give customers more for their money by including load balancing, firewalls and redundant networks,


In November 2012, Microsoft announced it plans to provide big compute capabilities in Windows Azure when it debuts its first virtualized InfiniBand RDMA (remote direct memory access) network, for those organizations looking to do HPC analyses. Another headwind ProfitBricks could face is that InfiniBand is pricier than other IaaS options.

- Time Warner/NaviSite : 
NaviSite is splitting the difference between colocation and managed services with pure cloud computing,


Over the past two years, NaviSite has shifted its focus from colocation to Infrastructure as a Service (IaaS),

- CloudSigma : flexible/give customer control
CloudSigma, an IaaS provider based in Zurich, Switzerland, was founded in 2009 as an alternative to public cloud giants, as a way to combine the flexibility and scalability of public cloud with the control customers were used to with managed hosting — a mission statement that could appeal to cloud-wary enterprises. Facilitate a more collaborative, flexible relationship between public cloud providers and customers.


CloudSigma's solution is really somewhere between managed services and pure cloud computing, Some enterprises will find that more desirable considering they are not giving up total control.”

[PaaS]
- EMC/VMware
Pivotal : combines cloud application development and big data analytics properties into a 1,400-person “virtual organization” within EMC, led by former VMware CEO Paul Maritz.

在今年3月份的VMware投资者大会上，VMware和EMC联合宣布将成立合资公司Pivotal，由EMC控股，EMC持有VMware约80%股份。Pivotal将由两家公司的数据分析和云应用资产合并而成，并且VMware公司前任CEO Paul Maritz出任Pivotal公司CEO，Pat Gelsinger接管VMware公司，David Goulden继续掌管EMC。北京时间2013年4月26日，Pivotal公司正式宣布成立，开始作为一个独立的实体运营。同时披露Pivotal One新一代PaaS计划，Pivotal One将是第一个集成新的分片式数据(Data Fabric)、现代编程框架、云便携性和遗留系统支持的平台。

EMC、VMware、Pivotal三家公司对详细业务进行了布局，其中EMC从事基础设施，VMware中间层，Pivotal则定位在顶端应用。Pivotal还宣布了Pivotal One计划，该计划旨在打造新一代企业PaaS平台。其中，Greenplum和Cloud Foundry是新公司的核心资产。Cloud Foundry是VMware推出的开源PaaS平台，Greenplum专注在大数据分析应用。Cetas作为大数据初创公司，也是Pivotal业务的补充。

这三家公司被业界称为“铁三角”，而且Pivotal之后，EMC系的企业级市场战略更加完善。三家公司有了明确分工，Pivotal定位在下一代云计算和大数据应用市场，VMware致力于软件定义数据中心，而EMC则专注在信息基础架构。

- CloudBees: Java PaaS and continuous delivery

Unlike other recently launched Platform as a Service vendors, CloudBees is focused on Java, JRails and Grails, rather than trying to be all things to all customers in terms of languages.

Instead, CloudBees offers vertical integration of the software process, pulling in continuous integration services through its Jenkins plugin, and it boasts partnerships with a number of ecosystem partners, including PaperTrail for log sequencing and New Relic for monitoring. Code is committed in CloudBees, tested by Jenkins and then deployed automatically to the CloudBees platform without the user having to worry about the care and feeding of each system along the way.

“Having Jenkins set up was a huge advantage for us, and allowed more time to work on core products,” said Mario Cruz, chief technology officer (CTO) and co-founder of Choose Digital, which switched to CloudBees from Heroku about a year ago.
“CloudBees is as much of a must-have to us as GitHub,” said Adrian Cole, founder of the jclouds open source project, which gets free space on CloudBees and has collaborated on the Jenkins plugin. The Jenkins plugin is also used by Google App Engine.

- Engine Yard : reduce vender lock-in concerns
Platform as a Service (PaaS) allows companies to quickly spin up applications to stay competitive in cutthroat verticals. But the ugly phrase “vendor lock-in” still looms over development teams’ heads. Engine Yard is hoping to change that.

By adding a new architecture to its Engine Yard Cloud, the San Francisco-based PaaS provider hopes to free developers from the need to re-architect apps for the platform. Its Infrastructure Abstraction layer gives developers the ability to deploy an application without worrying what other vendor(s) they are already running. And by partnering with various Infrastructure as a Service (IaaS) providers, Engine Yard hopes to minimize the dreaded fear of lock-in.

Part of its aim in minimizing developer strain might come from a growing emphasis on DevOps, which “represents the convergence and tighter integration of the build, plan, deploy and manage” cycle. makes the difference between the classic terms of ‘developer’ and ‘operations’ less meaningful.

- DotCloud : right price

dotCloud is another platform gunning for market leader Heroku, which launched in 2011, boasts tens of thousands of developer customers and is making inroads into enterprises through “guerilla” efforts — where developers at Fortune 1000 organizations lobby internally to use dotCloud for a specific project.

Designed for a new generation of cloud-native apps, with a focus on fast performance for Java development.

The Platform as a Service provider was also the only service available about a year ago that offered a stack that included Java, MongoDB, the Spring framework, the Dojo library and other specialized elements with an uptime service-level agreement (SLA). DotCloud’s SLA builds on Amazon Web Services IaaS’ 99.95% annual uptime guarantee.

- HeroKu: Market leader in PaaS

With a development community that has turned out well over 100,000 applications, along with the financial backing of Salesforce.com, which acquired Heroku in December 2010, Heroku seemed a lock to dominate the market among open source developers.

One reason for Heroku’s security is that it allows developers to build and deploy apps using not only Ruby, but also Node.js, Java, Python and Scala. Another reason is the array of management and fault-tolerant tools built into the base platform.

Late last year, Heroku delivered a couple of products that were received favorably, including Heroku Enterprise for Java, a service that allows IT shops to build and run applications in the cloud quickly. The product lets cloud providers gravitate to a continuous delivery model sooner rather than later. Additionally, the company, which introduced the idea of add-ons for PaaS providers, released version 2.0 of its Heroku Add-Ons Marketplace. The new release better enables business processes for cloud services providers, something that has gained more focus among customers.

Because Ruby has proven to be the popular programming language among developers creating social and mobile apps — the hottest app development segments at the moment – the Heroku cloud market position appears to be further strengthened.

Terremark- Acquired by Verizon

Manny Medina born in Cuba setup Terremark in 1980's, a real estate company constructing office buildings. During the dot-com era more and more of his buildings were leased to computer data centers and the company morphed into an information technology services company itself starting with the NAP of the Americas,^[2] Terremark's flagship facility, in downtown Miami. 

NAP of the Americas one of the most significant telecommunications projects in the world. The facility was the first purpose-built, carrier-neutral network access point (NAP) and is the only facility of its kind specifically designed to link Latin America with the rest of the world. The building is a 750,000 square foot, purpose-built datacenter. The equipment floors are 32 feet above sea level. The building is designed to withstand a Category 5 hurricane with approximately 19 million pounds of concrete roof ballast. It has 7 inch thick steel reinforced concrete exterior panels.^[3] NAP of the Americas carries 95% of the data traffic between North and South America.

Verizon Communications on January 27, 2011 announced it would buy Terremark Worldwide for $19 a share, in a deal valued at $1.4 billion. The deal for Verizon is a play on the rapid growth in cloud computing, an area where the company has been lagging behind competitors like AT&T.^[4] Terremark will now become a wholly owned subsidiary under Verizon, keeping its name, its employees and headquarters in Miami. 

Terremark now operates 13 data centers throughout the world. Medina will receive about $83 million out of the Verizon deal from his 4.4 million shares, almost 7 percent of the company.^[5] On May 10, 2011 Manny Medina received the Jay Malina Award from the Beacon Council. They stated that Terremark was Miami's leading force in the technology industry.^[6] Verizon completed its acquisition of Terremark on April 12, 2011.^[7] Medina left the company at the time of the takeover and Terremark has gone through two presidents in one year. Currently three high ranking executives are running the business.

Wiki page: http://en.wikipedia.org/wiki/Terremark

Cloud Computing - Amazon Web Service

1, From Wiki page: https://en.wikipedia.org/wiki/Cloud_computing

After the dot-com bubble, Amazon played a key role in all the development of cloud computing by modernizing their data centers, which, like most computer networks, were using as little as 10% of their capacity at any one time, just to leave room for occasional spikes. 

Having found that the new cloud architecture resulted in significant internal efficiency improvements whereby small, fast-moving "two-pizza teams" (teams small enough to feed with two pizzas) could add new features faster and more easily, Amazon initiated a new product development effort to provide cloud computing to external customers, and launched Amazon Web Services (AWS) on a utility computing basis in 2006.^[11][12]

In early 2008, Eucalyptus became the first open-source, AWS API-compatible platform for deploying private clouds. In early 2008, OpenNebula, enhanced in the RESERVOIR European Commission-funded project, became the first open-source software for deploying private and hybrid clouds, and for the federation of clouds.^[13] In the same year, efforts were focused on providing quality of service guarantees (as required by real-time interactive applications) to cloud-based infrastructures, in the framework of the IRMOS European Commission-funded project, resulting to a real-time cloud environment.^[14] 


2, From Wiki page: https://en.wikipedia.org/wiki/Amazon_Web_Services

Officially launched in 2006,^[3] Amazon Web Services provide online services for other web sites or client-side applications. Most of these services are not exposed directly to end users, but instead offer functionality that other developers can use in their applications. Amazon Web Services’ offerings are accessed over HTTP, using REST and SOAP protocols. All services are billed based on usage, but how usage is measured for billing varies from service to service.

In late 2003, Chris Pinkham and Benjamin Black presented a paper proposing the company could build and sell a set of services based on the experience of building and operating the infrastructure for Amazon.com.^[4] The first AWS service launched for public usage was Simple Queue Service in November 2004.^[5] Amazon EC2 was built by a team in Cape Town, South Africa under Pinkham and lead developer Chris Brown.^[6]

3, List of AWS products[edit]

Compute[edit]

• Amazon Elastic Compute Cloud (EC2) provides scalable virtual private servers using Xen.
• Amazon Elastic MapReduce (EMR) allows businesses, researchers, data analysts, and developers to easily and cheaply process vast amounts of data. It uses a hosted Hadoop framework running on the web-scale infrastructure ofEC2 and Amazon S3.

Networking[edit]

• Amazon Route 53 provides a highly available and scalable Domain Name System (DNS) web service.
• Amazon Virtual Private Cloud (VPC) creates a logically isolated set of Amazon EC2 instances which can be connected to an existing network using a VPN connection.
• AWS Direct Connect provides dedicated network connections into AWS data centers, providing faster and cheaper data throughput.

Content Delivery[edit]

• Amazon CloudFront, a content delivery network (CDN) for distributing objects to so-called "edge locations" near the requester.

Storage & Content Delivery[edit]

• Amazon Simple Storage Service (S3) provides Web Service based storage.
• Amazon Glacier, Provides a very low cost long-term storage option (when compared to its S3 service). High redundancy and availability, but low-frequent access times. Ideal for archiving data.
• AWS Storage Gateway, an iSCSI block storage virtual appliance with cloud-based backup.
• Amazon Elastic Block Store (EBS) provides persistent block-level storage volumes for EC2.
• AWS Import/Export, accelerates moving large amounts of data into and out of AWS using portable storage devices for transport.

Database[edit]

• Amazon DynamoDB provides a scalable, low-latency NoSQL online Database Service backed by SSDs.
• Amazon ElastiCache provides in-memory caching for web applications. This is Amazon's implementation of Memcached.
• Amazon Relational Database Service (RDS) provides a scalable database server with MySQL, Informix,^[15] Oracle, and SQL Server support.
• Amazon Redshift provides petabyte-scale data warehousing with column-based storage and multi-node compute.
• Amazon SimpleDB, allows developers to run queries on structured data. It operates in concert with EC2 and S3 to provide "the core functionality of a database."
• AWS Data Pipeline, provides reliable service for data transfer between different AWS compute and storage services(e.g. Amazon S3, Amazon RDS, Amazon DynamoDB, Amazon EMR.). In other words this service is simply a data-driven workload management system, which provides a simple management APIs to managing and monitoring of data-driven workloads in cloud applications.

Deployment[edit]

• Amazon CloudFormation provides a file based interface for provisioning other AWS resources.
• AWS Elastic Beanstalk provides quick deployment and management of applications in the cloud.
• AWS OpsWorks for configuration of EC2 services using Chef.

Management[edit]

• Amazon Identity and Access Management (IAM), an implicit service, the authentication infrastructure used to authenticate access to the various services.
• Amazon CloudWatch, provides monitoring for AWS cloud resources and applications, starting with EC2.
• AWS Management Console (AWS Console), A web-based point and click interface to manage and monitor the Amazon infrastructure suite including (but not limited to) EC2, EBS, S3, SQS, Amazon Elastic MapReduce, andAmazon CloudFront. Amazon also makes available a mobile application for the Android which has support for some of the management features from the console.

App Services[edit]

• Amazon CloudSearch provides basic full text search and indexing of textual content.
• Amazon DevPay, currently in limited beta version, is a billing and account management system for applications that developers have built atop Amazon Web Services.
• Amazon Elastic Transcoder (ETS) provides video transcoding of S3 hosted videos, marketed primarily as a way to convert source files into mobile-ready versions.
• Amazon Flexible Payments Service (FPS) provides an interface for micropayments.
• Amazon Simple Email Service (SES) provides bulk and transactional email sending.
• Amazon Simple Queue Service (SQS) provides a hosted message queue for web applications.
• Amazon Simple Notification Service (SNS) provides a hosted multiprotocol "push" messaging for applications.
• Amazon Simple Workflow (SWF) is a workflow service for building scalable, resilient applications.

AWS 101 by Amazon: http://www.youtube.com/watch?v=uwhHSosGF0w (Jul 2013)

Introduction to AWS:http://www.youtube.com/watch?v=CaJCmoGIW24 (Oct 2011)

Good talk about the benefits to move to AWS cloud: http://www.youtube.com/watch?v=yMJ75k9X5_8 (Jan 2013)


Auto-Scaling: Launch configuration, group, policy(metrics based or schedule based)

Why adopting cloud computing?
- Variable expense: replace capital expenditure with variable expense: CapEx -> OpEx
- Economics of scale: lower variable expense than companies can achieve themselves; the more usage, the cheaper cost per unit;
- Elastic capacity: No need to guess capacity requirements and over-provision: demand/time chart;
- Speed and agility: infrastructure in minutes not weeks;
- focus on business: not undifferentiated heavy IT lifting ( Data Center, Power, Cooling, Cabling, Racks, Servers, Storage/Labour, buy and install new hardware, setup and config new software)
- Global Reach, Go global in minutes and reach a global audience