Archive of entries from 2012

12/31/2012

OpEx

Operating Expenditures

IaaS, SaaS, PaaS costs are considered operating expenses as they are paid on a recurring basis in flexible increments.  Infrastructure stacks (hardware, virtualization, OS, database, associated licensing and support) are rolled into this monthly cost to the business.

 

12/28/2012

CapEx

Capital Expenditures

In the technology context, CapEx are the large investments in non-cloud infrastructure.  Buying bulky servers, networking equipment, licensing, and software which needs to be configured, procured, and maintained by IT staff.  Cloud computing provides an alternative to these clumsy CapEx purchases with on-demand, utilitized, virtual servers and equipment.

 

12/27/2012

Mashup Web App

Application which combines content and/or functionality from multiple sources

Mashups aggregate and stitch together third-party data into one, interactive user interface.  The web application spreads roots across the Web, utilizing data and functionality retrieved from data sources that lay outside its platform boundaries.  Several APIs may be utilized.

For more information, see this article at IBM developerWorks.

 

12/26/2012

Perimeter Control 

oversight of network intrusion at the edge of a company’s infrastructure.  In the cloud context, normally a port and IP traffic-blocking firewall

Perimeter control is the first line of defense against intruders, autonomously allowing only necessary traffic through to service-providing servers and devices.  The term is normally referred to in the IT context as the firewalls and edge devices used to allow and block traffic into the DMZ or internal network but may be expanded to include physical security such as card readers, physical data center keys, security guards, etc.

 

12/21/2012

Utilitization

The act of taking a product or service and applying a financial model to it which charges for metered usage.

Whether termed the cloud or utility computing, a new economics model exists where computing infrastructure is moved to specialized providers who can use their scale to cost-effectively deliver this service.  Compare this shift away from local data centers to that of the electricity model of the early 20th century.  It was expensive for organizations to create their own electricity plant to power a factory.  The same as those factories, cloud customers are now taking advantage of utilitized cloud offerings so they can focus on product innovation and the core business, rather than underlying storage, networking infrastructure, hardware failover, etc.

 

12/20/2012

TCO

Total Cost of Ownership – total direct and indirect costs of a product or system

Because pricing structures of on-premises infrastructure and IaaS differ so much, accounting analysis may be necessary to realize the bottom line differences.  CapEx, IT OpEx, and business-level OpEx should be examined.  The TCO model should also extend over the expected lifecycle of applications and hardware.

 

12/19/2012

VPC

Virtual Private Cloud – an internal cloud which is exists within a public cloud and the internal network can access its resources directly

Example, Amazon VPC allows customers to configure a VPN to connect internal infrastructure to servers, storage, and other services within Amazon Web Services.  Employees at the local office can log into servers in a similar fashion to servers at the local datacenter.

 

12/18/2012

Internal Cloud

infrastructure which supports services within an organization’s scope

Internal cloud is used loosely around the industry and may mean any service offered to an organization’s employees by the IT staff.  Virtualization and flexibility are commonly associated with the internal cloud but the term does not necessarily describe any single technology.

aka Private Cloud

 

12/17/2012

Hybrid Cloud

a networking and server environment which includes both local, private infrastructure and a direct connection to one or more public clouds

Companies may create a VPN connection from their local datacenter to a public cloud like Rackspace or Amazon Web Services.  Servers can be spun up at the provider, joined to domain, or connected to locally.  Direct access to storage and other services is also possible from the company office.

 

12/14/2012

HaaS

Hardware as a Service, physical equipment is provided in an environment over the internet by the provider

Companies requiring dedicated, physical hardware may use HaaS.  Their servers and network equipment is housed at a hosting provider but the customer never touches or sees the physical machines.  Clients with compliance requirements or those which require high resource utilization benefit from the segmentation of dedicated hardware or the direct access to resources.  Rackspace is an industry leader in HaaS.

 

12/13/2012

Consumption-Based Pricing Model

customers are charged based on the amount of service or product they consume

A cloud service provider may charge per gigabyte for storage rather than a static, X dollars per month with a storage limit.  Another example: a consumer pays for the number of hours their servers are turned on rather than buying the server.

 

12/12/2012

CloudStorming

connecting multiple cloud providers within a single infrastructure

An example of cloudstorming: a company uses both Rackspace and AWS to host pieces of their infrastructure.  VPN connections are set up to both using AWS VPC or Rackspace RackConnect®.  A Microsoft Active Directory topology may segment a site for each environment.

 

12/11/2012

Vendor Lock-In

situation where an organization is unable to change between vendors because their data, processes or infrastructure are inextricably linked with one particular provider.

This has always been a problem in the technology realm and is nothing new for Iaas, PaaS, or SaaS.  Most cloud providers are built on a standard set of products and processes which are portable.  Industry professionals encourage Open Standards and several organizations have been formed to advocate a set of industry-wide principles.  The costs of re-architecting scripts, load balancers, servers, network, and other items need to be considered.

 

12/10/2012

Amalgamation

In business, a merging of business units, offerings, or workflows into one.

Multiple cloud computing providers are working toward a standardized, portable set of development and infrastructure models.  Amalgamation results in industry communities where people with similar philosophies about cloud architecture and open-source software can openly collaborate together.  A good example is in the OpenStack participating companies.

 

12/7/2012

EDA, Event-Driven Architecture

A framework that orchestrates and reacts to any significant, identifiable occurrence for infrastructure, platform, or software

An EDA consists of event creators and consumers.  Creators, usually some type of middleware event manager, is aware the event has occurred.  Consumers receive notification of an event from a creator and react, process, or may be simply affected by the event.

Example, a consumer buys coffee from a distributor.  The coffee inventory level lowers below a threshold.  The distributor’s system treats this state change as an event which is published to various applications within the architecture.  The website may remove the item from being displayed and an inventory system initiates the process to order more coffee.

 

12/6/2012

URI, Uniform Resource Identifier

The internet space is inhabited by many points of content.  A URI is a unique way to identify those text pages, images, videos, or sound clips.

The most common form of URI is a web URL.  However, URIs can be more generally defined and can describe:

  • The specific server or computer that a resource is housed in
  • The specific filename of the piece of content
  • The path or mechanism used to access a resource

 

12/5/2012

Internet of Things

concept that all objects and people in daily life are uniquely identifiable and virtual representations of these unique IDs can be interconnected

If the world is filled with interconnected, sensor-laden devices theoretically builds a smarter planet (like the IBM commercials).  An example application is with stoplights, bus schedules, and benches.  If a bus stop bench reports 6 sitting people, stop lights may be altered to speed the bus on route.  The cloud is seen as a step towards full-scale “Internet of Things.”  The vast scalability means computing can grow and shrink to meet demand, including all of those uniquely identifiable objects.

 

12/4/2012

Intercloud 

theory that each cloud can extend itself to other clouds to increase resource availability and geographic redundancy

Intercloud is based on the grid concept that each, single cloud does not have unlimited geographic footprint or infinite resources.  If one cloud saturates its own infrastructure, it can still satisfy requests for service from clients.  The scenario would processing, storage, or other resources can be allocated from other clouds.  The commodity is easily transferable and may eventually create a more open market between vendors.

 

12/3/2012

Utility Computing

computing resources packaged as a metered service, delivered on-demand

The “aaS” in IaaS, PaaS, and SaaS means the related resources are delivered as a service, in a utility computing fashion.  Adoption of this delivery model has dramatically increased between 2010 and 2012.

IDC estimates that 24% of new business software purchases will be delivered as a service by the year 2015.

IDC – Market Analysis Perspective: Worldwide SaaS & Cloud Services, 2011: New Models for Delivering Software

 

11/30/2012

Cloud Oriented Architecture, COA

applications and services running in the cloud can be interrelated with other objects in the environment

Jeff Bezos of Amazon coined the abstract term which has a wide-ranging scope.  COA addresses the architectural constraints of traditional IT vs the flexibility of the cloud.  Adoption of COA means it is easier for everything be interrelated, the “internet of things” concept where computers, stop lights, buildings, park benches can all be uniquely identified and coordinate with other, related objects.

 

11/29/2012

Cloud Broker

An entity which creates connections to multiple cloud service providers on disparate infrastructures, acting as a liaison on the customer’s behalf

Enterprises gain a unified cloud infrastructure without the risk of lock-in or loss of services in the event of a system failure.  Examples of cloud brokers vary in size and application.  The Defense Information Systems Agency has been designated as the US Department of Defense’s enterprise cloud provider.  Business-level companies like Infosys and Rightscale offer unification solutions to enterprises and the SMB market.

 

11/28/2012

Cloud Gaming

Game playing online that allows streaming of games to a computer, play from any device, or transferable settings between consoles

Zynga allowed users to break into the cloud gaming industry with hits like Farmville and Words with Friends.  These games were accessible from any computer and now, any device.

aka: gaming on demand

 

11/27/2012

Cloud Engineering 

application of engineering methods to bring a systematic approach to the high-level concerns of standardization, jurisdiction, and commercialization of cloud computing systems

Cloud engineering is the high-level view which focuses on all disciplines; software, infrastructure and platform.  As cloud computing has become commercially viable, oversight has been necessary to maintain a set of standards.  These standards make it possible for these services to be interchangeable and comparable in the open market.

 

11/26/2012

Location Independence

computing service delivery attribute in which end users can run software applications from any location

Similar to and often coupled with device independence, location independence means an end user can run services from work, home, coffee shops, cell phones or tablets.

 

11/21/2012

Device Independence

computing service delivery attribute in which an application can function on devices regardless of the local hardware or operating system

Device Independence is a key strength of the SaaS delivery model.  Users access the same software on their PC, phone, or tablet.  Services are configured on cloud infrastructure are often optimized to be device independent.

Related terms: Hardware Agnostic, Platform Agnostic

 

11/20/2012

Client-Server Model

a computing model which contains providers, servers or resources, and requesters, called clients.  Clients and servers often communicate over a network.

This model has changed in recent years with the advent of cloud computing.  During the golden years of IT, clients and servers were contained within local LANs where employees worked within the confines of the building walls.  Now, cloud delivery models make services available from anywhere on multiple devices.  Clients connect over the internet rather than on the local LAN.

 

11/19/2012

SDK

Software Development Kit – a set of development tools which allow for the creation of software applications, built for a specific framework, hardware platform, computer system,  operating system, or similar platform.

How is an SDK different from an API?  An SDK is tooling for building something while an API is the interface.  An analogy would be electrical wiring in your house.  An API would be the power connections between an appliance and the wall while an SDK may be a kit to build a custom appliance.  -a build-it-yourself toaster or something similar.

Examples of an SDKs include the iPhone app developmentSDK, Android SDK, or Xbox Development Kit.

 

11/16/2012

Multi-Instance Architecture

separate instances (hardware, virtual, or software) are configured for different client organizations

Contrasted with multitenancy where a single instance runs on a server and serves multiple client organizations.

An example: a development firm creates a web application for its clients.  In multi-instance architecture, each client has its own server on which the software is housed.  Multitenancy is a single server and web app which serves multiple clients.  Each client has their own login and customized software but all back end infrastructure is combined.

 

11/15/2012

REST-based API

REpresentational State Transfer Application Programming Interface – for the web, a stack of software architecture for distributed systems which follows basic guidelines:

  1. Each resource has an ID.  the web service has a base URI (http://website.com/webapp)
  2. The media type is a common, supported hypertext standard (many times XML but can be others)
  3. The API must be hypertext driven
  4. A standard set of API operations is supported within HTTP methods (GET, PUT, DELETE)

REST is not a “standard” published by any group or organization.  The purest meaning of REST is just the architectural style.  This style can be applied to HTTP and the web.

Alternatives are SOAP, RPC, CORBA and WSDL.  The main differentiators for REST is agility of design, ease of implementation, better cache support, reduced network traffic, and an overall lightweight approach.

 

11/14/2012

Ubiquitous Computing

technologies are omnipresent, present in everyday objects and activities.

Any device, from your wallet, to thermostat, to cars can be embedded with technology to connect the device to a network of other devices.

aka: ubicomp, pervasive computing, ambient intelligence

 

11/13/2012

Multitenancy

a single instance of software serves multiple customers or roles at one time

Multitenant applications are generally more efficient than the traditional client-server hosting model, where the server to customer ratio is 1:1.  The other key attribute that makes cloud multitenant applications attractive is their scalability.  Application data, computing and other resources are virtually partitioned.  Examples of multitenant applications include Facebook, Google Apps like Gmail on the consumer side and Salesforce or LinkedIn in the business realm.

 

11/12/2012

Skeuomorphism

a design technique that imitates functionally necessary design elements in original products but have now become ornamental

An example of a skeuomorph is the leather binder in Apple’s iBooks or a rolodex icon signifying a contact list.  The same can be seen in cloud infrastructure.  Virtual network switches may be represented by a picture of a network card, save functionality is a 3.5″ floppy disk, and virtual machines are shown as a physical server box.

 

11/9/2012

Pedagogical

of, relating to teaching or education, ranging from human development to skills acquisition

How is this related to the cloud?  Benchmarks in the technical side of IT departments are typically wrapped in a certification or continuing education process.  While the cloud is new, many people are learning its core concepts through hands on trial and error.  The industry lends itself to this type of learning since there is no large investment necessary.  Pedagogically, the concepts are easy to study and teach.  The hands on availability makes it easy to apply these theories in a practical, usable way.

 

11/8/2012

APIaaS

API as a Service – publicly-available service platform which enables procedures and subroutines to be centrally hosted and available from anywhere

When a company wants to provide developers with a programming interface to their software, they may provide APIaaS.  The differentiator for APIaaS from API is the platform is centrally hosted, provided publicly where procedures and subroutines can be executed remotely.  Programmers do not need to download the API and run procedures on their local platform.  APIaaS may be rolled into a PaaS or SaaS solution, providing developers with test, dev, and production environments.

Examples of APIaaS include the Twitter and Facebook website plugins.

 

11/7/2012

DaaS

Desktop as a Service – a centrally hosted computing desktop environment, separate from a physical location or machine

The end user is delivered the information system environment (email, file servers, internal apps) or the environment itself (a full desktop) to a remote client device.  The client device may use an entirely different hardware architecture from that used by the projected desktop environment, and may also be based upon an entirely different operating system.  A user may log in from a thin client locally, through a web portal over the internet, or from mobile devices.

aka: Virtual Desktop Infrastructure (VDI), client virtualization

 

11/6/2012

DaaS

Data as a Service – data is presented from a centralized location to applications or users, regardless of their location or device

An example of DaaS is Xignite who provides financial market data on demand via their API.  Another is Urban Mapping, a geography data service that allows real-estate companies to embed data in their own sites and apps.

 

11/5/2012

TEaaS

Test Environment as a Service – a delivery model in which a software development platform is centrally hosted, scalable, and available on-demand

Developers can code utilizing TEaaS’s hardware, network, and operating system.  These development solutions are then presented to end users to confirm functionality before merging to production.  When software solutions are presented to end users, these are sometimes referred to as pre-release versions or beta.

 

11/2/2012

ITaaS

IT as a Service – all-encompassing business model describing an IT department which utilizes cloud as a part of their service offering.

IaaS, PaaS, SaaS, STaaS, SECaaS, DaaS, TEaaS, APIaaS

This IT methodology means IT departments are flexible and procedures are set to administer systems from anywhere.  Distributed business units are supported by a distributed IT department.

 

11/1/2012

SaaS

Software as a Service – software delivery model where application and associated data are centrally hosted on the cloud.  Users do not manage the infrastructure or platform on which the application is running

Examples of SaaS include Microsoft Office 365, Google Apps, and Salesforce.com applications.  Rather than selling a local copy and license of Microsoft Office for $100-350, a SaaS model provides the software online for $6-20 per month.  All documents, email, spreadsheets, and other files are hosted by Microsoft and are accessed independently of device or location.

 

10/31/2012

PaaS

Platform as a Service – cloud computing service model which provides a centrally hosted software development solution stack, offering the facilities to deploy applications from anywhere without buying underlying hardware and software

Examples of PaaS include Windows Azure and Google App Engine.  Platform is the layer between the software and the infrastructure, providing developers with specialized APIs in their specific programming segment.  For example, a .NET developer looking for a place to host a web application, someone who uses SQL for his database backend and IIS for web hosting, may look to Microsoft Azure.

 

10/30/2012

IaaS

Infrastructure as a Service – a service model of cloud computing which centrally hosts infrastructure, servers, network, VMs, storage, load balancers, security devices

Examples of IaaS include Amazon Web Services and Rackspace offerings.  IaaS is the building block on which PaaS and SaaS are built.  It replaces traditional on-site servers and networking.  The solutions are elastic, administered from anywhere, and on-demand, pay for what you use.

IaaS, PaaS, SaaS cloud stack:

IaaS, PaaS, SaaS

 

10/29/2012

Scale Down

The process of provisioning smaller instances to match demand

For example, a customer normally runs their web application on 2 Medium Amazon EC2 instances behind a load balancer.  If traffic to those apps dramatically increases, those 2 Medium Amazon EC2 instances may resize down to small size.

aka: scale vertically

  • 3.75 GB memory
  • 2 EC2 Compute Unit (1 virtual core with 2 EC2 Compute Unit)
  • 410 GB instance storage
  • 32-bit or 64-bit platform
  • I/O Performance: Moderate
  • EBS-Optimized Available: No
  • API name: m1.medium
  • 1.7 GB memory
  • 1 EC2 Compute Unit (1 virtual core with 1 EC2 Compute Unit)
  • 160 GB instance storage
  • 32-bit or 64-bit platform
  • I/O Performance: Moderate
  • EBS-Optimized Available: No
  • API name: m1.small

AWS Scale Down

 

10/26/2012

Scale In

decreasing the number of resources to match demand

Example, a customer normally runs their web application on 4 Medium Amazon EC2 instances behind a load balancer.  During the night, load decreases to minimal levels.  The customer may scale in their infrastructure, changing from 4 to 2 instances.

aka: scale horizontally

AWS Scale In

 

10/25/2012

Scale Up

provisioning larger resources to match demand

For example, a customer normally runs their web application on 2 Medium Amazon EC2 instances behind a load balancer.  If traffic to those apps dramatically increases, those 2 Medium Amazon EC2 instances may resize up to large size.

aka: scale vertically

  • 3.75 GB memory
  • 2 EC2 Compute Unit (1 virtual core with 2 EC2 Compute Unit)
  • 410 GB instance storage
  • 32-bit or 64-bit platform
  • I/O Performance: Moderate
  • EBS-Optimized Available: No
  • API name: m1.medium
  • 7.5 GB memory
  • 4 EC2 Compute Units (2 virtual cores with 2 EC2 Compute Units each)
  • 850 GB instance storage
  • 64-bit platform
  • I/O Performance: High
  • EBS-Optimized Available: 500 Mbps
  • API name: m1.large

 

Amazon EC2 Scale Up

 

10/24/2012

Scale Out

provisioning a higher quanitity of resources to match demand

For example, a customer normally runs their web application on 2 Amazon EC2 instances behind a load balancer.  If traffic to those apps dramatically increases, 2 additional EC2 instances may be deployed, providing 4 to match demand and minimize latency demand.

aka: scale horizontally

 

Amazon EC2 Scale Out

 

10/23/2012

Elasticity

In the cloud computing context, the ability to deploy application resources nearly instantly and then quickly resize those resources as workload changes.

Companies like Rackspace and Amazon Web Services pitch elasticity as one of the core differentiators from traditional IT models.  If an application reaches its peak demand at noon every weekday, 20 servers may be running behind the load balancer, responding to client web requests, while during the night, when demand drops, only 5 hosts remain up.  In a single server scenario, the host may run at 8 CPUs during the day and scale down to 2 at night.  The cloud customer only pays for the resources used.

 

10/22/2012

HPC Clusters

High Performance Computing clusters are engineered to provide high-performance network and computational capabilities, allowing applications to get the performance required for tightly coupled, node-to-node communication.

HPC Clusters allow scientists,  engineers, and business analysts to solve complex problems.  Typically, these applications must wait to access shared clusters, supercomputers, or acquire expensive hardware systems.  Using cloud instances, customers can spin up low cost, elastic resources and pay for only what they use.

 

10/19/2012

Commoditization

The process by which goods and services with distinguishable economic value become comparable commodities in the eyes of the market.

Commoditization may be a movement from a differentiated to an undifferentiated price or from monopolistic to healthy competition.  In the IT context, computing power is in the process of being commoditized as it moves from specific vendors at local datacenters to cloud providers where services are interchangeable.  Rackspace’s Cloud Server offering can easily be compared to that of an EC2 Instance at Amazon Web Services.

 

10/18/2012

Sustainability

The responsible management of resource use.  Generally, the human capacity for caretaking of environmental, economic and social dimensions.

While cloud computing is assumed to be “green”, viable evidence is yet to be produced that demonstrates the public cloud to support that thought.  Energy efficiency may be gained at scale when providers practice energy-aware scheduling.  Certain geographic areas may be more efficient due to their capacity for natural cooling due to climate or access to renewable electricity.

An outside the box thought has been raised by Microsoft research.   The proposal is to place distributed servers in homes and offices, using the server as a primary heat source.

 

10/17/2012

Software Defined Networking

A networking architecture in which the control plane of network equipment is separate from the data plane, meaning external access to the administrative interfaces is available where they were previously proprietary or inaccessible

The definition of SDN has varied over time.  The current emerging sentiment focuses on the programmatic access to network equipment.  Cloud providers like AWS and Rackspace provide customers APIs to load balancers, firewalls, and routing protocols.

 

10/16/2012

OpEx

Operating expenditures.  -Cloud computing costs are considered OpEx as they are paid on a recurring basis in small, manageable, flexible increments.  Infrastructure stacks (hardware, virtualization, OS, database, associated licensing and support) are “utilitized” and billed as they are used.

 

10/15/2012

CapEx

Capital expenditures.  In the cloud computing context, CapEx are the large investments in non-cloud infrastructure.  Buying bulky (physically and on cash flows) servers, networking equipment, licensing, and software which needs to be configured, procured, and maintained by IT staff.  Cloud computing provides an alternative to these clumsy CapEx purchases with on-demand, utilitized, virtual servers and equipment.

 

10/12/2012

OSSM

Acronym used to describe the cloud: On-demand, Self-service, Scalable, Measurable

Servers are on-demand, meaning images and instances are configured to best practices for their specific role and are ready to be deployed nearly instantaneously.  Customers have self-service control of their resources.  Computing capacity grows/multiplies and contracts/shrinks to closely match demand, meaning servers are scalable.  Measurable, real-time metering and reporting lets customers know exactly what is being utilized.

*via Dave Nielsen at CloudCamp

 

10/11/2012

DevOps

A job role or title which includes both software programming (dev) and server/network administration (ops)

The cloud is demanding skillsets be diverse and two old IT silos, development and operations, have combined into one job description.  DevOps individuals generally bridge the gap between these two silos and can translate software engineers’ requirements to the server and network teams.

 

10/10/2012

Democratization

In the cloud computing context, the transition which allows any individual or business to access storage, computing power, bandwidth, operating systems, applications and databases any time and anywhere at a pricing model which charges for only what is used.

Enterprise-level infrastructure previously only available to those organizations who could afford it is available to anyone with an internet connection.  There is no need to invest thousands of dollars to obtain the hardware, software, and maintenance necessary to run high performance applications.  The pre-cloud computing paradigms, total cost of ownership, staffing, and server room administration, that held back so many businesses have been disrupted.

 

10/9/2012

Peak Load Capacity

In the cloud context, the potential ability of an infrastructure to perform, yield, or withstand peak demand

Peak load is the time of highest resource demand from an application.  Traditionally, equipment needed to be requisitioned to handle peak load but those resources became underutilized during other timeframes throughout the day.  With the cloud, admins do not need to engineer the environment for the highest load levels.  Resources can be scaled up or down to match demand.

 

10/8/2012

Hardware Abstraction

sets of routines in an operating system or applications which provide a platform to access underlying hardware resources

Early computing systems did not have a hardware abstraction layer (HAL) and software developers needed to know how to communicate with each hardware device.  With HALs, software instructs the operating system what it needs from a device and the OS translates those instructions.

If you have ever run into driver issues on a Microsoft Windows Server during a P2V, sysprep, or when moving between virtualization platforms, a common problem is with the HAL-related DLLs.  For example, a Standard PC may use hal.dll while an ACPI Multiprocessor PC utilizes halmacpi.dll.  Device Manager does not allow for changing between these HALs and a new build of the OS is required.

 

10/5/2012

Competitive Differentiation

products, processes, people, business ideals which set a business apart from the competition

Competitive differentiation applies to the cloud when a company’s technology strategy gives them an advantage over competitors.  Cloud adoption flattens the playing field.

For example, Laura’s Insurance Company runs virtual servers, making Laura’s datacenter run more efficiently at less cost.  Curt’s Insurance Company runs all physical servers which are expensive to run and maintain.  Curt’s business runs with higher expenses and must charge more for the product.  When both companies adopt the cloud as a part of the IT infrastructure, operating expenses can be leveled and competitive differentiation is equalized.

While cloud offerings are the same to Laura as they are Curt, efficiencies are present and can be realized.  Having a mature cloud strategy gives IT staff the comfort level and experience to take advantage of these efficiencies.  Early adopters have seen their infrastructures develop into auto-scaling, elastic systems which consistently track to the demand of the business.

 

10/4/2012

Big Data

a term generally used to describe data sets so large and complex they become awkward to work with using single instance, on-hand database management tools.

Big data is difficult to work with using relational databases, requiring instead enormous, scalable software running in parallel on tens, hundreds, or thousands of servers.  Examples of platforms on which to run big data analytics include Apache Hadoop (the de facto standard), HPCC Systems, Azure Table Storage, Backtype Storm (Storm focuses on realtime processing.  Twitter acquired Backtype in Summer 2011). Amazon Web Services Elastic Map Reduce is Hadoop running on AWS.

 

10/3/2012

Rationalized Access

More commonly known as role-based access, ensures access is needs based, rather than all users having permissions to all services the organization uses

For example, DBAs may only have access to specific MS SQL databases.  Security groups may have rights to specific cloud storage buckets.  For enterprise-level Microsoft environments, permissions can be assigned based on Active Directory users and groups.  In cloud infrastructure, Amazon Web Services has a well-developed role-based system for accessing AWS resources.

 

10/2/2012

Compute Cycle

a unit of measurement in cloud environments which represents an aggregation of resource usage.  PaaS applications commonly bill based on compute cycles which is calculated using processor as the core metric.

This measurement varies from one provider to the next.  A common benchmark is that 10,000 compute cycles is roughly equivalent to a modern processor.  Rackspace defines as 10k CC = one 2.8GHz server processor capacity although disk IO may also figure into the final number.  Memory and bandwidth generally function independently of Compute Cycles.

 

10/1/2012

Grid Computing

a distributed system running from multiple locations or domains.  Grid computing typically handles large, non-interactive workloads which involve large databases, high volumes of files, or CPU-extensive processes

A grid member computer is typically connected by a network interface instead of the high-speed computer bus in typical supercomputers.  A common application of grid computing is in CPU scavenging where unused resources of network participants, home computers, servers, etc, run instruction cycles while the user is away at night or at lunch.

Folding@Home is a Stanford University project used to understand proteins, protein misfolds, and associated diseases.  The project allows users to download and run client software to increase simulation speed by providing unused computing power from their local machine.

aka virtual supercomputers

 

9/28/2012

Access Control

the ability to monitor, verify, and restrict connections between people, assets, or information.

In the castle analogy, access control is the guard, asking for identification or a password, verifying you have access to certain areas of the internal environment.  In the IT context, access control may include subnet control, VLAN segmentation, file security, and other controls in place which check an individual’s credentials and group membership.

Access Control relates to permitting or denying access, an interactive process, while perimeter control is intrusion prevention.

 

9/27/2012

Perimeter Control

physical control of network intrusion through mechanical means.  In the cloud context, normally a port and IP-blocking firewall

Perimeter control is the castle walls, a first line of defense against intruders.  The term is normally referred to in the IT context as the firewalls and edge devices used to allow and block traffic into the DMZ or internal network but may be expanded to include physical security such as card readers, physical datacenter keys, security guards, etc.

 

9/26/2012

“Not Built Here” mentality

Internal IT’s belief that technology solutions should always be built in house and that any solution that is obtained from third party provider is sub-optimal when compared to something built and deployed within the organization.

 

9/25/2012

Opportunity Cost

a basic economic concept which factors the costs related to the choices not made by someone.  Opportunity costs are not restricted to direct monetary gain.  Lost time, morale, output, and other indirect factors should also be considered.

This is an important concept in IT decision making as there may be no direct cost associated with maintaining the status quo of an onsite datacenter, physical servers, software licensing, etc.  Studies show that roughly 80% of IT effort and expenditure is used on processes that do not create value for the organization but go towards maintaining that status quo.  If a portion of this 80% can be redirected to core business processes and applications, a move to the cloud is a benefit which may not be realized unless that optimal option is considered.

 

9/24/2012

Utilitization

The act of taking a product or service and applying a financial model to it which charges for metered usage.

Whether termed the cloud or utility computing, a new economics model exists where computing infrastructure is moved to specialized providers who can use their scale to cost-effectively deliver this service.  Compare this shift away from local data centers to that of the electricity model of the early 20th century.  It was expensive for organizations to create their own electricity plant to power a factory.  The same as those factories, cloud customers are now taking advantage of utilitized cloud offerings so they can focus on product innovation and the core business, rather than underlying storage, networking infrastructure, hardware failover, etc.

 

9/21/2012

TCO

Total Cost of Ownership – total direct and indirect costs of a product or system

Because pricing structures of on-premises infrastructure and IaaS differ so much, accounting analysis may be necessary to realize the bottom line differences.  CapEx, IT OpEx, and business-level OpEx should be examined.  The TCO model should also extend over the expected lifecycle of applications and hardware.

 

9/20/2012

Paravirtualization

Paravirtualization is a method which allows multiple operating systems and components to run on shared hardware.  Paravirtualization’s management module operates with a closer relationship to the underlying hardware within the virtual machine’s operating system built to optimize that underlying hardware.  This is where paravirtual techniques differ from full virtualization where all components (machine CPU, memory, disk, BIOS, network card) in a VM is emulated.  Performance is improved when some critical tasks are executed relocated within host hardware as opposed to being fully run in the virtual domain.

Paravirtual components will often be offered to virtual machines.  These components will allow tasks to be run within the underlying hardware than other adapters.  The virtual machine operating system must support this virtual hardware through its driver sets.

 

9/19/2012

Virtualization

Virtualization is a technique which allows multiple operating systems, storage, network, or other component platforms to run within shared hardware at the same time.  Multiple virtual machines (VMs) may run on a single server.  The VM operating system and end users see separate servers running segmented roles.  From the hardware’s perspective, the VMs are merely sets of files which share resources like CPU and memory.

 

9/18/2012

Vendor Lock-In

A situation in which an organization is unable to move vendors because their data, processes and infrastructure are inextricably linked with one particular provider.

This has always been a problem in the technology realm and is nothing new for Iaas, PaaS, or SaaS.  Most cloud providers are built on a standard set of products and processes which are portable.  Industry professionals encourage Open Standards and several organizations have been formed to advocate a set of industry-wide principles.  The costs of re-architecting load balancers and auto scale groups need to be considered.  A majority of pricing models have been stable for a long time but situations like the Google App Engine price increase late last year could arise.

 

9/17/2012

Amalgamation

In business, a merging of business units, undertakings, or workflows into one.  In cloud computing, the work of multiple organizations toward a standardized, portable set of development and infrastructure models.  An industry community where people with similar philosophies about cloud architecture and open-source software can openly collaborate together.

We believe in open source, open design, open development and an open community that is fully transparent.  -as in the open standards cloud, common APIs.  A good example is in the OpenStack participating companies.

 

9/14/2012

Intercloud 

The idea that each cloud is an extension of the internet and may be interconnected to other clouds to increase resource allocation and geographic topology

Intercloud is based on the grid concept that each, single cloud does not have unlimited geographic footprint or infinite resources.  If a cloud saturates its own infrastructure, it can still satisfy requests for service from clients.  The scenario would processing, storage, or other resources can be allocated from other clouds.  The commodity is easily transferable and may eventually create a more open market between vendors.

 

9/13/2012

ITaaS

IT as a Service – an all-encompassing business model describing an in-house IT department utilizing cloud as a part of their service offering.

IaaS, PaaS, SaaS, STaaS, SECaaS, DaaS, TEaaS, APIaaS

This IT methodology means IT departments are flexible and procedures are set to administer systems from anywhere.  Distributed business units are supported by a distributed IT department.

 

9/12/2012

TEaaS

Test Environment as a Service – a delivery model in which a software development platform is centrally hosted, an on-demand test environment

Developers can code utilizing TEaaS’s hardware, network, and operating system.  These development solutions are then presented to end users to confirm functionality before merging to production.  When software solutions are presented to end users, these are sometimes referred to as pre-release versions or beta.

 

9/11/2012

DaaS

Data as a Service – data is presented from a centralized location, offering it to different systems, applications or users, regardless of their location or device

An example of DaaS is Xignite who provides financial market data on demand via their API.  Another is Urban Mapping, a geography data service that allows real-estate companies to embed data in their own sites and apps.

 

9/10/2012

DaaS

Desktop as a Service – a personal computing desktop environment separate from a physical location or machine

The end user is delivered the information system environment (email, file servers, internal apps) or the environment itself (a full desktop) to a remote client device.  The client device may use an entirely different hardware architecture from that used by the projected desktop environment, and may also be based upon an entirely different operating system.  A user may log in from a thin client locally, through a web portal over the internet, or from mobile devices.

aka: Virtual Desktop Infrastructure (VDI), client virtualization

 

9/7/2012

Commoditization

The process by which goods and services have distinguishable economic value in terms of uniqueness or brand become comparable commodities in the eyes of the market.

This may be a movement from a differentiated to an undifferentiated price or from monopolistic to healthy competition.  In the IT context, computing power is in the process of being commoditized as it moves from specific vendors at local datacenters to cloud providers where services are interchangeable.

 

9/6/2012

Autonomic Computing

self-managing characteristics of computing resources, those which automatically adjust to demand and change while hiding intrinsic complexities of underlying architecture from operators and users

 

9/5/2012

Utility Computing

computing resources packaged as a metered service, delivered on-demand

The “aaS” in IaaS, PaaS, and SaaS means the related resources are delivered as a service, in a utility computing fashion.  Adoption of this delivery model has dramatically increased between 2010 and 2012.

IDC estimates that 24% of new business software purchases will be delivered as a service by the year 2015.

IDC – Market Analysis Perspective: Worldwide SaaS & Cloud Services, 2011: New Models for Delivering Software

 

9/4/2012

Grid Computing

a distributed system running from multiple locations or domains.  Grid computing typically handles large, non-interactive workloads which involve large databases, high volumes of files, or CPU-extensive processes

A grid member computer is typically connected by a network interface instead of the high-speed computer bus in typical supercomputers.  A common application of grid computing is in CPU scavenging where unused resources of network participants, home computers, servers, etc, run instruction cycles while the user is away at night or at lunch.

Folding@Home is a Stanford University project used to understand proteins, protein misfolds, and associated diseases.  The project allows users to download and run client software to increase simulation speed by providing unused computing power from their local machine.

aka virtual supercomputers

 

8/31/2012

Cloud Gaming

Game playing online that allows streaming of games to a computer, play from any device, or transferable settings between consoles

Zynga allowed users to break into the cloud gaming industry with hits like Farmville and Words with Friends.  These games were accessible from any computer and now, any device.

aka: gaming on demand

 

8/30/2012

Location Independence

a computing service delivery attribute in which end users are able to run a software application from any location

Similar to and often coupled with device independence, location independence means an end user can run services from work, home, coffee shops, cell phones or tablets.

 

8/29/2012

Device Independence

a computing service delivery attribute in which a software application is able to function on multiple devices regardless of the local hardware or operating system

Device Independence is a key strength of the SaaS delivery model.  Users access the same software on their PC, phone, or tablet.  Services are configured on cloud infrastructure are often optimized to be device independent.

Related terms: Hardware Agnostic, Platform Agnostic

 

8/28/2012

Client-Server Model

a computing model which contains providers, servers or resources, and requesters, called clients.  Clients and servers often communicate over a network.

This model has changed in recent years with the advent of cloud computing.  During the golden years of IT, clients and servers were contained within local LANs where employees worked within the confines of the building walls.  Now, cloud delivery models make services available from anywhere on multiple devices.  Clients connect over the internet rather than on the local LAN.

 

8/27/2012

Middleware

computer software which provides services and interface to developers beyond those available from a base operating system

The following functions may be performed by middleware:

  • hide distribution of processing, network, or storage
  • provide hardware agnostic functionality to the software layer
  • supply common commands and functions to the software layer to avoid additional coding
  • provide standardized interfaces for developers so applications can interoperate and be easily ported

The term can be placed rather generally.  In the cloud computing context, middleware may give developers and system integrators better insight and administration of their underlying infrastructure.

 

Finishing an API-focused week…

8/24/2012

OCCI

Open Cloud Computing Interface – a set of open, community-driven specifications for public cloud resource interactions

OCCI provides a protocol and common API components for IaaS based services.  The current OCCI specification set consists of three documents, with future releases planned to include additional specifications.

http://occi-wg.org/

 

8/23/2012

SDK

Software Development Kit – a set of development tools which allow for the creation of software applications, built for a specific framework, hardware platform, computer system,  operating system, or similar platform.

How is an SDK different from an API?  An SDK is tooling for building something while an API is the interface.  An analogy would be electrical wiring in your house.  An API would be the power connections between an appliance and the wall while an SDK may be a kit to build a custom appliance.  -a build-it-yourself toaster or something similar.

Examples of an SDKs include the iPhone app developmentSDK, Android SDK, or Xbox Development Kit.

 

8/22/2012

APIaaS

API as a Service – publicly-available service platform which enables procedures and subroutines to be run from a hosted, centralized environment

When a company wants to provide developers with a programming interface to their software, they may provide APIaaS.  The differentiator for APIaaS from API is the platform is centrally hosted, provided publicly where procedures and subroutines can be executed remotely.  Programmers do not need to download the API and run procedures on their local platform.  APIaaS may be rolled into a PaaS or SaaS solution, providing developers with test, dev, and production environments.

Examples of APIaaS include the Twitter and Facebook website plugins.

 

8/21/2012

REST-based API

REpresentational State Transfer Application Programming Interface – for the web, a stack of software architecture for distributed systems which follows basic guidelines:

  1. Each resource has an ID.  the web service has a base URI (http://website.com/webapp)
  2. The media type is a common, supported hypertext standard (many times XML but can be others)
  3. The API must be hypertext driven
  4. A standard set of API operations is supported within HTTP methods (GET, PUT, DELETE)

REST is not a “standard” published by any group or organization.  The purest meaning of REST is just the architectural style.  This style can be applied to HTTP and the web.

Alternatives are SOAP, RPC, CORBA and WSDL.  The main differentiators for REST is agility of design, ease of implementation, better cache support, reduced network traffic, and an overall lightweight approach.

 

8/20/2012

API

Application Programming Interface – set of routines, protocols, and tools for building software applications.

APIs are many times vendor-specific, providing the building blocks for a developer to put all the pieces together in a final product: desktop, web, mobile apps.  Cloud provider APIs allow web developers to programmatically provision, scale and configure their infrastructure/platform.

 

8/17/2012

Ubiquitous Computing

a computing model in which technologies are omnipresent, being integrated into everyday objects and activities.

Any device, from your wallet, to thermostat, to cars can be embedded with technology to connect the device to a network of other devices.

aka: ubicomp, pervasive computing, ambient intelligence

8/16/2012

Multitenancy

a single instance of software serves multiple, diverse customers at the same time

Multitenant applications are much more efficient than the traditional client-server hosting model, where the server to customer ratio is 1:1.  The other key attribute that makes cloud multitenant applications attractive is their scalability.  Examples of multitenant applications include Facebook, Google Apps like Gmail on the consumer side and Salesforce or LinkedIn in the business realm.

 

IaaS, PaaS, SaaS cloud stack:

IaaS, PaaS, SaaS

8/15/2012

IaaS

Infrastructure as a Service – a service model of cloud computing which centrally hosts virtual machines, servers, storage, load balancers, network

Examples of IaaS include Amazon Web Services and Rackspace offerings.  IaaS is the building block on which PaaS and SaaS are built.  It replaces traditional on-site servers and networking.  The solutions are elastic, administered from anywhere, and on-demand, pay for what you use.

 

8/14/2012

PaaS

Platform as a Service – a service model of cloud computing which provides a software development solution stack as a service, offering the facilities to deploy applications from anywhere without buying underlying hardware and software

Examples of PaaS include Windows Azure and Google App Engine.  PaaS is the layer between the software and the infrastructure, providing developers with specialized APIs in their specific programming segment.  For example, a .NET developer looking for a place to host a web application, someone who uses SQL for his database backend and IIS for web hosting, may look to Microsoft Azure.

 

8/13/2012

SaaS

Software as a Service – a software delivery model where application and associated data are centrally hosted on the cloud.  Users do not manage the infrastructure or platform on which the application is running

Examples of SaaS include Microsoft Office 365, Google Apps, and Salesforce.com applications.  Rather than selling a local copy and license of Microsoft Office for $100-350, a SaaS model provides the software online for $6-20 per month.  All documents, email, spreadsheets, and other files are hosted by Microsoft and are accessed independently of device or location.

 

8/10/2012

Opportunity Cost

a basic economic concept which factors the costs related to the choices not made by someone.  Opportunity costs are not restricted to direct monetary gain.  Lost time, morale, output, and other indirect factors should also be considered.

This is an important concept in IT decision making as there may be no direct cost associated with maintaining the status quo of an onsite datacenter, physical servers, software licensing, etc.  Studies show that roughly 80% of IT effort and expenditure is used on processes that do not create value for the organization but go towards maintaining that status quo.  If a portion of this 80% can be redirected to core business processes and applications, a move to the cloud is a benefit which may not be realized unless that optimal option is considered.

 

8/9/2012

TCO

Total Cost of Ownership – total direct and indirect costs of a product or system

Because pricing structures of on-premises infrastructure and IaaS differ so much, accounting analysis may be necessary to realize the bottom line differences.  CapEx, IT OpEx, and business-level OpEx should be examined.  The TCO model should also extend over the expected lifecycle of applications and hardware.

 

8/8/2012

OSSM

Acronym used to describe the cloud: On-demand, Self-service, Scalable, Measurable *

Servers are on-demand, meaning images and instances are configured to best practices for their specific role and are ready to be deployed nearly instantaneously.  Customers have self-service control of their resources.  Computing capacity grows/multiplies and contracts/shrinks to closely match demand, meaning servers are scalable.  Measurable, real-time metering and reporting lets customers know exactly what is being utilized.

*via Dave Nielsen at CloudCamp

 

8/7/2012

Commoditization

The process by which goods and services have distinguishable economic value in terms of uniqueness or brand become comparable commodities in the eyes of the market.

This may be a movement from a differentiated to an undifferentiated price or from monopolistic to healthy competition.  In the IT context, computing power is in the process of being commoditized as it moves from specific vendors at local datacenters to cloud providers where services are interchangeable.

 

8/6/2012

“Not Built Here” mentality

Internal IT’s belief that technology solutions should always be built in house and that any solution that is obtained from third party provider is sub-optimal when compared to something built and deployed within the organization.

 

8/3/2012

Democratization

In the cloud computing context, the transition which allows any individual or business to access storage, computing power, bandwidth, operating systems, applications and databases any time and anywhere at a pricing model which charges for only what is used.

Enterprise-level infrastructure previously only available to those organizations who could afford it is available to anyone with an internet connection.  There is no need to invest thousands of dollars to obtain the hardware, software, and maintenance necessary to run high performance applications.  The pre-cloud computing paradigms, total cost of ownership, staffing, and server room administration, that held back so many businesses have been disrupted.

 

8/2/2012

Big Data

a term generally used to describe data sets so large and complex they become awkward to work with using single instance, on-hand database management tools.

Big data is difficult to work with using relational databases.  Instead enormous, scalable analytics software running in parallel on tens, hundreds, or thousands of servers is required.  Examples of platforms on which to run big data analytics include Apache Hadoop (the de facto standard), HPCC Systems, Azure Table Storage, Backtype Storm (Storm focuses on realtime processing.  Twitter acquired Backtype in Summer 2011). Amazon Web Services Elastic Map Reduce is Hadoop running on AWS.

 

8/1/2012

Utilitization

The act of taking a product or service and applying a financial model to it which charges for metered usage.

Whether termed the cloud or utility computing, a new economics model exists where computing infrastructure is moved to specialized providers who can use their scale to cost-effectively deliver this service.  Compare this shift away from local data centers to that of the electricity model of the early 20th century.  It was expensive for organizations to create their own electricity plant to power a factory.  The same as those factories, cloud customers are now taking advantage of utilitized cloud offerings so they can focus on product innovation and the core business, rather than underlying storage, networking infrastructure, hardware failover, etc.

 

7/31/2012

OpEx

Operating expenditures Cloud computing costs are considered operating expenses as they are paid on a recurring basis in small, manageable, flexible increments.  Infrastructure stacks (hardware, virtualization, OS, database, associated licensing and support) are rolled into this monthly cost to the business.

 

7/30/2012

CapEx

Capital expenditures For the cloud computing context, CapEx are the large investments in non-cloud infrastructure.  Buying bulky (physically and on cash flows) servers, networking equipment, licensing, and software which needs to be configured, procured, and maintained by IT staff.  Cloud computing provides an alternative to these clumsy CapEx purchases with on-demand, utilitized, virtual servers and equipment.

 

7/27/2012

Rationalized Access

More commonly known as role-based access, ensures that access is needs based, rather than all users having permissions to all services the organization uses

For example, DBAs may only have access to MySQL.  Specific groups may have security rights to specific cloud storage buckets.  For enterprise-level Microsoft environments, permissions can be assigned based on Active Directory users and groups.  Amazon Web Services has a well-developed role-based system for accessing AWS resources.

 

7/26/2012

Amalgamation

In business, a merging of business units, undertakings, or workflows into one.  In cloud computing, the work of multiple organizations toward a standardized, portable set of development and infrastructure models.An industry community where people with similar philosophies about cloud architecture and open-source software can openly collaborate together.

We believe in open source, open design, open development and an open community that is fully transparent.  -as in the open standards cloud, common APIs.  A good example is in the OpenStack participating companies.

 

7/25/2012

Vendor Lock-In

A situation in which an organization is unable to move vendors because their data, processes and infrastructure are inextricably linked with one particular provider.

This has always been a problem in the technology realm and is nothing new for Iaas, PaaS, or SaaS.  Most cloud providers are built on a standard set of products and processes which are portable.  Industry professionals encourage Open Standards and several organizations have been formed to advocate a set of industry-wide principles.  The costs of re-architecting load balancers and auto scale groups need to be considered.  A majority of pricing models have been stable for a long time but situations like the Google App Engine price increase late last year could arise.

 

7/24/2012

Sunk Costs

retrospective costs that have already been incurred.  -a generally well known business term, for the cloud, capital expenditures towards existing, internal IT hardware, software, and licensing which is still undepreciated

Classic economics states that sunk costs should not affect decision making.  However, it is commonplace for stakeholders to deny their organization an opportunity to innovate because of a past, bad buying decision. While internal IT is considered CapEx, IaaS and cloud resources are OpEx, meaning 2 things; 1. a migration from internal to cloud may incur some sunk costs for past capital expenditures and 2. sunk costs are limited in a cloud project, turn off the servers, network, and other infrastructure at any time.

 

7/23/2012

Fungible

adjective, [fuhn-juh-buhl] – the properties of a good or commodity as to be freely exchangeable or replaceable, for another of like properties

When the hardware that powers servers is indistinguishable and interchangeable, it becomes a fungible commodity.  This interchangeable property means that applications and data are open and portable between providers.  Servers can run on cloud infrastructure the way an appliance runs on electricity or a car runs on gasoline.