Datex Intelligent
Data Center Services
Heralding A New Era in Advanced Internet Operations
Technology
Background
The Internet was originally built as a collection of small networks. First the
Universities connected their networks together, then connected with the big
telecommunication companies, and then others. As more and more networks began connecting,
these NAPs as they're commonly called (network access points) became the points through
which Internet traffic was routed. There are several major and hundreds and hundreds of
minor NAPs through which Internet traffic may be routed.
The network architecture of the Internet, by design, is
essentially dumb. Nodes are connected to nodes. Each node and NAP by itself is essentially
"dumb" -- essentially handoff points, however the chaos created by the
combination of the dumb nodes makes the Internet essentially invulnerable to failing.
Denial of service attacks against backbone carriers, human errors at the nodes,
misconfigurations, etc.. all occur on a daily basis, however the Internet as we know it
remains functioning. This chaotic foundation of the Internet has essentially freed the
Internet from single points of failure. It's this chaos that Datex is replicating and
adding "intelligence" to a layer above the Internet - at the Data Center
level.
The old way of outsourcing data center services
As the Internet began to commercialize back in 1994, companies found that outsourcing
their Internet Infrastructure was the way to go. It relieved them of technical headaches
to manage a 24/7 Internet operations, they lacked the internal expertise to manage sites
that were heavily trafficked, they could not keep recruiting the top technical talent
necessary to maintain secure Internet presences, and more. Thus sprang the emergence of
companies like Exodus Communications - the major facilities companies who specialized in
housing servers.
A big problem emerged for Companies who colocated their
Operations at these facilities however. Rather than take advantage of the Internet's
chaotic (and 100% reliable) nature, companies like Exodus themselves became points of
failure for Internet Operations. Housing your Internet Infrastructure in one facility, on
a single network, or with a single facilities player exposes your Infrastructure to a high
degree of vulnerability.

The diagram above shows what happens in traditionally
outsourced Internet Operations. Your Internet Operations is sitting on the
"chaotic" Internet - which is highly invulnerable to failure. However by
colocating Internet Operations at a facilities player, like Frontier Communications
(Globalcenter) you have again inserted a high degree of risk (the vulnerability of a
single point of failure) into your Internet Operations.
If the Frontier network were hit with a denial of service
attack or some human error leaves routers misconfigured at the facilities, your servers
can become unreachable to your customers. It is estimated today that even very good
facilities like Frontier's or Exodus' go "down" or unreachable for an average of
30 minutes per day due to common, everyday occurrences like network maintenance,
restarting routers, human errors, etc..
Case in point: AboveNet's Outage
The highly publicized outage at AboveNet in February 2000 is case in point. AboveNet's
claims of offering a bullet-proof 100% available web presence were thwarted. AboveNet's
network was the target of a hacker who flooded their network with incoming traffic (a
denial of service attack). The result was that while AboveNet went through common
restoration and troubleshooting procedures, all of AboveNet and their customers were
offline for 3-4 hours. This included all of their locations on the west coast, east coast,
etc.. This spurred many of AboveNet's customers who depend on an always available Internet
presence to look for ways in which they could spread their Internet Infrastructures across
multiple networks, data centers, and facilities to add redundancy and Insurance against
having this type of occurrence lead to lost sales ever again.
The ideal solution
The ideal Internet Infrastructure for companies with mission critical Internet Operations
is one that is deployed in a chaotic fashion across all nodes on the Internet with instant
replicas of their content (both static and dynamic) at all locations. This would give that
company ubiquitous presence and essentially guarantee them 100% uptime. No single hack
attack or denial of service attack on a network would put a halt to their business.
Many hosts today offer geographic clustering services to
attempt to replicate this chaos, however even clustering is lacking because 1) it does not
provide for the instant replication of your static and dynamic content 2) it does not
monitor the performance of services on your server, such as checking to see whether your
database transactions are being completed, whether your Vignette or Broadvision
applications are running properly, etc.
To achieve true optimization of Internet Operations, a
company must monitor the performance levels to the servers (internet traffic), the
applications that reside on those servers (making sure transactions to Oracle or DB2
databases are being logged for example), and the servers themselves are functioning
maximally (high uptime, no disk errors, port conditions, etc.) and be able to proactively
provision traffic to the most optimally performing server in the Company's Internet
Infrastructure.
Datex Intelligent Infrastructure Technology
This is where Datex enters the picture. Datex inserts intelligence into Internet
Infrastructures using its proprietary technology that has been developed over the course
of several years. Datex has found that to achieve 100% uptime, companies need their
"presence" spread across a number of networks and facilities and geographic
locations, and those extra presences should be "hot" backups of the content
that's at the main location. Datex's technology allows companies to decide whether to use
those alternate locations as active servers, or passive servers waiting for calamity at
the main data centers. Datex's technology automatically monitors the Internet
Infrastructure from the perspective of the end-user to ensure that all services and
applications are running optimally and that latency times for web surfers are minimal.

Datex essentially combines the "chaotic" nature
of the Internet with its Intelligent Data Center technology to ensure maximum reliability,
hot swappable standby "Insurance Servers" and maximum end-user performance for
hosted infrastructures. If your Company offers Service Level Agreements to your end-users,
if you cannot afford any downtime, and cannot afford to miss transactions, then you need
Datex's Intelligent Infrastructure Technology. Best of all, our technology can be used in
conjunction with your current Internet Infrastructure.
Contact a Datex sales
engineer to see how our high availability solutions can help your Internet Infrastructure
- and help you keep disaster from striking.
|