44
instartlogic.com
Copyright © 2013 Instart Logic, Inc. All rights reserved.
4
WHITE PAPER
With the rise of social media, most publishers have adopted personalization to make apps more fun and more
relevant, and to make it easier to turn apps into extensions of a user’s social graph. At the same time, publishers
have built a greater degree of dynamism into next-generation web apps. This ensures that users not only have
contextually relevant experiences, but also that the app remains fresh at all times, encouraging users to spend
more time there. All of this is made more possible by powerful browsers, which increasingly take on computa-
tional tasks formerly handled by backend web servers.
To meet user demands for richer applications and to take advantage of the technology advances of HTML5, web
publishers have continued to make their applications larger and more complex. These applications, however,
tend to diminish responsiveness because the initial application load requires more bandwidth and faster down-
load speeds. The conflict between richer applications and demands for improved application responsiveness is
particularly pronounced for delivery over congested wireless networks, including 3G, 4G, and WiFi.
Wireless connections tend to lag behind wired connections in terms of speed and capacity for a number of rea-
sons. Wireless connections all have in common shared and limited spectrum and hard-to-control competition for
resources (bandwidth and download capacity) by users in the same geographical area. Unlike wired connections,
wireless connection speeds can vary greatly from one moment to the next. This is problematic for web designers,
because they must build an application that delivers acceptable user experiences regardless of the connection
speed or local state of the wireless connection. Additionally, problems in the architecture of basic web proto-
cols (HTTP) result in further latencies, because they hinder multi-threading and cause cascading slowdowns as
browsers wait for specific requests to be fulfilled in turn before additional pieces of an application can be down-
loaded to the device.
Despite the best efforts by even the most technically advanced web publishers, the shared nature of wireless
connections and the unpredictable loads experienced in any region of a mobile infrastructure can lead to bottle-
necks and degraded responsiveness. Users encountering slow-loading web sites and applications find it to be
a poor user experience. Slow delivery, too, can negatively impact a website’s ranking in search engines: Google,
for example, factors the responsiveness of a site into its PageRank algorithms. Slower delivery also results in
fewer pages of a website being indexed by Google’s crawlers, which operate on timed intervals. To resolve these
conflicts, web publishers have resorted to mobile versions of their applications that use highly compressed
images and “slimmed-down” user experiences with fewer dynamic and visual components. While this improves
responsiveness and reduces time-to-interaction, it also means mobile versions are less attractive, less interac-
tive, and potentially less lucrative than full-featured desktop or laptop versions.
Why Existing Technology Approaches Can’t Resolve This Conflict
Legacy technologies for web application delivery have struggled to deliver on the twin requirements of respon-
siveness and richness.
Content Delivery Networks
The first efforts to combat poor web performance were Content Delivery Networks (CDNs). CDNs are cloud-
based services that improve responsiveness and time-to-interactivity by keeping exact copies of frequently-ac-
cessed images, scripts, and other website components cached closer to users in ISP edge networks. CDNs were
originally developed over 10 years ago when latency and routing problems in the core of the Internet were the
primary bottlenecks to fast website delivery. Back then, the Internet backbone was much slower and less reliable
than it is today.
VB.NET PDF - Annotate PDF with WPF PDF Viewer for VB.NET Text box. Click to add a text box to specific location on PDF page. Line color and fill can be set in properties. Copyright © <2000-2016> by <RasterEdge.com>.
convert pdf to pdf form fillable; convert word to pdf fillable form online
39
instartlogic.com
Copyright © 2013 Instart Logic, Inc. All rights reserved.
5
WHITE PAPER
The largest CDNs have thousands of Points of Presence (PoPs) around the globe and maintain highly complex
network topologies, a remnant of the old, slow, unreliable Internet. This infrastructure is extremely expensive to
maintain and operate, a cost that is generally passed on to customers. Many newer CDNs have since emerged
with fewer PoPs because the modern Internet core is so much faster, better connected, and reliable. They can
provide the same coverage with less than two dozen properly located PoPs at Internet peering points and deliver
performance that is indistinguishable to end users from that delivered by the larger legacy CDNs.
Because CDNs were conceived of and designed for an era when most websites were static and users were on
wired last mile networks, they were not designed to address the new challenges of highly personalized, graph-
ically rich, dynamic websites and applications accessed over wireless last mile networks. Personalization and
the increasing reliance of applications on a user’s social graph reduce the efficacy of keeping copies of images
cached close to the edge of the Internet but before last mile networks. Two users may be on the same website in
the same room, but because that website imports images from their social graph or reflects their personal tastes,
the experience they are served may be completely different. In a similar manner, modern users expect websites
and web apps they visit to change frequently. Caching, by definition, relies on relatively infrequent changes to
content stored on the edge of the network.
Application Delivery Networks
Application Delivery Networks (ADNs) evolved to alleviate some of the failings of CDNs when web sites became
much more personalized. This technology adds network acceleration to existing CDN technology to help speed
the delivery of personalized web site components, which can’t be cached. However, ADNs are only an incremen-
tal enhancement to the static caching capability provided by CDNs. Much like CDNs, this acceleration approach
stops at the edge of the Internet. And now that the bottleneck has shifted past the edge to the wireless last mile,
ADNs don’t properly address the challenges of modern web application delivery in this wireless world.
Front End Optimization
The latest web performance approach has been Front End Optimization (FEO) solutions which automatically ap-
ply web performance best practices to web pages by making heavy modifications of the core content and code
of websites and web apps. This includes invasive modifications of HTML, CSS, JavaScript and images – often
in ways that the original author of the code never foresaw or considered while developing their application. FEO
solutions can result in broken websites and service interruptions.
Since most sophisticated web publishers have already implemented web performance best practices in house
and have likely taken them several steps further than FEO solutions can provide, these solutions are of less value.
FEO solutions tend to make the most sense for smaller companies that lack the in-house expertise or time to
focus on web performance best practices.
Some CDNs are now attempting to address modern delivery challenges by incorporating FEO technologies into
their offerings and overly focusing on reducing the quality of images or sending multiple copies of images to im-
prove performance, at the cost of user experience, and passing higher data use charges on to end users.
38
instartlogic.com
Copyright © 2013 Instart Logic, Inc. All rights reserved.
6
WHITE PAPER
A New Approach: The Web Application Streaming Network
In the early days of the Internet, users that wanted to view a video online would need to download the entire file
before it would start to play. This was a laborious process. Streaming video sped up this process by eliminating
the need to download the entire video file before it could be watched. With the advent of streaming video, users
could almost immediately start watching a video after only a partial download of data, while the remainder con-
tinued to download in the background.
As with video downloads, previously users of Windows applications needed to download the entire application
up front before it could be installed and used. Then Windows Application Streaming came along. Like video
streaming, it eliminated the need to send all the data up front before an application became useable. Windows
application streaming works by sending the most important application data first and allowing users to start us-
ing Windows apps after downloading only a small portion of the data, while the rest continues to download in the
background. Just as with the streaming video approach, the user gets a faster experience, while still providing
the full capabilities of the application.
Instart Logic’s technology takes the lessons learned from streaming video and streaming desktop Windows
applications and applies them to web sites and applications with the world’s first Web Application Streaming
Network™. This approach allows users to interact with a web site or web application before the entire download
has completed. By establishing a cloud/client architecture linking the browser on the end user’s device to an
intelligent streaming service in the cloud, Instart Logic can determine which parts of a web application are most
important and deliver that information first. The best part is that Instart Logic does not require web publishers to
modify any code or for end users to download any plug-ins or applications. All that is required for a web publish-
er to use the Web Application Streaming Network is a simple change to DNS settings to direct end-user traffic
through the Instart Logic service. The technology works with standard browsers and does not require custom
plug-ins, custom browsers, or additional software. Instart Logic’s service is completely transparent to end users.
This technology resolves the conflict between responsiveness and richness and allows web publishers to de-
liver the most complex, image-rich, dynamically-generated, highly-personalized applications while maintaining
web performance that is radically faster than legacy CDN and ADN technologies can provide. In addition, Instart
Logic’s Web Application Streaming Network is a drop-in replacement for what CDNs or ADNs offer for Internet
middle mile acceleration. And while the Web Application Streaming Network was designed for the worst-case
scenario of mobile users on wireless networks, it delivers benefits for all users and devices.
A Deeper Dive into the Instart Logic Web Application Streaming Network
There are two unique components to the Web Application Streaming Network: the intelligent client-side
component called the HTML5 NanoVisor™ and the application-aware cloud service called the Personalized
AppSequencer™. The two work together to actively enable streaming of web sites and applications from web
publishers servers to standard web browsers.
36
instartlogic.com
Copyright © 2013 Instart Logic, Inc. All rights reserved.
7
WHITE PAPER
HTML5 NanoVisor
The HTML5 NanoVisor is a thin virtualization layer that runs in an end user’s browser, and sits between the
browser’s API and cache and the web site or application. The NanoVisor runs automatically when a web ap-
plication loads and is completely transparent to the user. It is written entirely in JavaScript, which is the key to
running transparently, without requiring any changes to the browser or application. It allows the Instart Logic
service to profile how a browser loads and executes a web application and then monitor which assets for that
site are requested and in what sequence. It then reports back the load profile to Instart Logic’s Personalized
AppSequencer. The HTML5 NanoVisor then enables standard browsers to receive a web site or application as a
stream from the Instart Logic cloud service.
Personalized AppSequencer
The Personalized AppSequencer is a distributed cloud-based service that performs the actual fragmentation
of web site or application assets, including images, HTML and Flash. The Personalized AppSequencer has a
deep understanding of the component parts of web pages and, from interacting with the HTML5 NanoVisor,
understands how they are loaded and executed in a browser. For instance, it understands the detailed structure
of image formats such as JPG, PNG, and WebP, it knows the difference between the unique and non-unique
portions of HTML, and it understands the detailed structure and format of Flash applications. The Personalized
AppSequencer automatically prepares application components for streaming.
True Fidelity Image Streaming™
The Personalized AppSequencer transcodes images, applying the most current lossless compression tech-
niques, and then allows them to be divided into fragments based on the underlying format support (such as
progressive display for JPEG images or interlacing for PNG images). This allows the system to send the first
portion of image data to allow quick initial display of the web site or application. The Personalized AppSequencer
then sends the remainder of the image quality information in the background soon after the initial page display,
providing a full fidelity experience. It allows users to interact with the entire web application faster and lets web
publishers load higher-quality images and content without diminishing performance.
Client Side
Intelligence
Deep Application
Understanding
Works with any Web
Infrastructure
Figure A: Instart Logic Key Components
44
instartlogic.com
Copyright © 2013 Instart Logic, Inc. All rights reserved.
8
WHITE PAPER
Dynamic HTML Streaming
The Personalized AppSequencer also understands the makeup of dynamic HTML pages and can identify sec-
tions of HTML that are the same across different users accessing the same page or application. Examples are
CSS and JavaScript files, header information and metadata, along with logos and other non-unique navigation
elements.
Normally for sites serving dynamic HTML, the backend web server has to generate unique HTML for each user.
This process, combined with the time required to deliver the HTML across the Internet, means the browser idles
after making the initial request. With Dynamic HTML streaming enabled, the Personalized AppSequencer takes
advantage of this idle time to send any non-unique portions of the HTML down to the browser from an Instart
Logic serving location close to the end user making the request.
The Personalized AppSequencer does this by looking at the HTML for the same URL for previous users’ re-
quests. If the content is dynamic HTML, it looks for portions that are the same across different users. This
information is then stored in the Instart Logic service. When the same page is requested the next time, the non-
unique HTML is sent immediately to the requesting browser while the dynamic content is requested from the
originating backend web server. As a result, the browser starts processing the non-unique sections and down-
loading components needed to render the page (for example JavaScript, CSS, fonts, and other items contained
in the non-unique sections) rather than sitting idle.
Flash Application Streaming
The Personalized AppSequencer also has a deep understanding of Flash games and applications. It can peer
into large SWF files and understand the component parts such as images, shapes, audio, video, and ActionScript
code. The Personalized AppSequencer can then fragment the SWF file so that, with a subset of data, the Flash
game or application can load and become interactive on the end user’s system while the remainder of the infor-
mation then comes down in the background after the user starts playing the game or using the application. Given
the large size of Flash games and applications, this has a dramatic effect on reducing the time to first interaction
for end users.
Putting It All Together
When the first user visits a page, Instart Logic will automatically deploy the NanoVisor to learn the load and exe-
cution profile of the specific URL. The Personalized AppSequencer will then automatically fragment the compo-
nent parts of the page and cache them in its globally distributed serving network.
After this first request, the AppSequencer gains sufficient intelligence about an application to effectively repri-
oritize asset resource requests; then the image streaming, dynamic HTML, or Flash application streaming tech-
nology takes effect. Most importantly, this allows users to begin interacting with the web experience as addi-
tional portions of the assets continue to stream down in the background. This capability dramatically cuts the
time-to-interaction for web publishers and makes their applications significantly more engaging. As a result of
this capability, Instart Logic customers see their users increasing time spent on site, returning more often, and
converting more transactions.
With this patented new cloud/client architecture, Instart Logic can also continue to add the ability to stream addi-
tional types of web components and further expand the capabilities of the Web Application Streaming Network.
For non-streamable components such as JSON, XML, JS and Fonts, the system will store and send the full
components.
19
instartlogic.com
Copyright © 2013 Instart Logic, Inc. All rights reserved.
9
WHITE PAPER
A Global, Enterprise Grade Network Built for the Real-Time Web
The Web Application Streaming Network is globally distributed and has serving sites in 30 locations around
the world and are next to every major Internet peering point. This coverage affords fast connectivity and glob-
al coverage near all major population centers and on all populated continents. Instart Logic’s Web Application
Streaming Network incorporates all the key capabilities of legacy CDNs in terms of network distribution, network
acceleration, and backend server offload and traffic spike protection, but was designed from scratch with a
unique web application streaming architecture to address modern web experience challenges.
To ensure total reliability, each tier of Instart Logic’s infrastructure – servers, cabinets, fiber connections, data-
bases, web servers — is fully redundant, with automatic failover capabilities between components and serving
locations. Within each location, Instart Logic has secured redundant peering arrangements with multiple tier 1
ISPs to ensure seamless global connectivity, even in the event of catastrophic provider failure. The system also
uses multiple levels of load balancing, with redundant local load balancing in each serving location, and global
load balancing across its distributed infrastructure. According to 3rd party measurements, the Instart Logic Web
Application Streaming Network has enjoyed 99.9999% uptime since launching officially in the summer of 2012.
Figure B: Instart Logic Global Network Locations
38
instartlogic.com
Copyright © 2013 Instart Logic, Inc. All rights reserved.
10
WHITE PAPER
Fast to Set Up, Easy to Manage, Easy to Measure
Web publishers wishing to use Instart Logic’s cloud service only need to make a few quick changes to their DNS
settings to flow traffic through the Instart Logic service. They do not need to change any code on their servers
or add new hardware or software to their web infrastructure, and their users do not need to download and install
any software or plug-ins to benefit from the Web Application Streaming Network. In addition to making migration
easier, Instart Logic allows web publishers to test the service on a portion of their live traffic before moving all
traffic over to it.
Web publishers can then easily manage their Instart Logic accounts by accessing the web-based customer
portal or using a REST API. The portal provides easy-to-use dashboards allowing customers to easily monitor
activity levels and measure performance globally and in specific countries. And to make it easy to measure the
amazing results from the service, web publishers can use third-party web application performance monitoring
tools, such as New Relic, Soasta mPulse, webpagetest.org and others, to benchmark before-and-after perfor-
mance gains with the service.
Conclusion
Instart Logic’s Web Application Streaming Network is a breakthrough technology that establishes a new par-
adigm for web performance by enabling web publishers to stream web applications directly to any standard
web browser. This technology revolutionizes the delivery of dynamic, personalized, geo-specific and image-rich
applications that have become the de facto standard for cutting-edge web publishers and a requirement for web
users with powerful, high-resolution tablets and smartphones. By assessing load profiles of web applications and
then intelligently dividing key assets (images, Flash, HTML) into smaller fragments, which are then streamed to
the device in order of priority, the Web Application Streaming Network allows users to interact with apps signifi-
cantly faster – particularly on devices loading apps over congested wireless networks. For web publishers, Instart
Logic’s streaming technology can dramatically reduce time-to-first-interaction and page load times, boosting rev-
enues and engagement while enabling them to deliver more personalized experiences that don’t sacrifice quality
in the interest of speed.
Because Instart Logic is a cloud technology with a unique cloud/client architecture, additional capabilities will
also continue to be added to the system, rendering this a “future-proof” solution for web publishers seeking
longer-term solutions for next-generation application delivery challenges. The Instart Logic Web Application
Streaming Network is a drop-in replacement for legacy CDN and ADN technologies, delivering superior perfor-
mance that drives superior user experiences.
Need more information?
Ready to get started?
DOWNLOAD RESOURCES
GET A FREE TRIAL
Documents you may be interested
Documents you may be interested
