Networking is becoming more central to computing than ever
before. With wireless becoming increasingly prevalent, and new
devices making mobile data more usable (and, therefore, more
relevant), companies are having to take more notice of the way that
they manage their digital communications.
The march of network convergence and the continuing encroachment
of IP into every part of the networking world also presents its own
challenges. Computer Weekly takes a look at the significant
networking trends and events in 2008.
Wimax out of the spotlight
Worldwide Interoperability for Microwave Access (Wimax) is like
unified communications, or public key infrastructure (PKI) before
that. Every year seems to be the one when it will take off, but it
never seems to happen.
In late 2007, Intel was touting 2008 as the high-speed wireless
standard's big year, but things have not panned out that way.
In the UK, Ofcom delayed a crucial 2GHz spectrum sale, which put
a spanner in the works for operators mulling widespread Wimax
services.
In the US, Sprint launched a mobile Wimax service in September,
and Clearwire already had a fixed wireless service, but a joint
venture between the two companies has been frustratingly slow to
emerge.
The two firms originally planned to get together on a
country-wide Wimax service in 2007, but cancelled the deal. "Sprint
had a massive plan for Wimax and it all got canned. That was a
major body blow," says Scott Morrison, a research vice-president at
Gartner.
Then the two companies reconvened, and have just officially
launched their joint venture, called Clear. The plan is to build a
nationwide Wimax network to rival AT&T's in 2009, but that will
require a large capital expenditure.
A number of companies including Intel, Comcast and Google, have
piled money into the $14bn initiative. Perhaps, once again, Wimax's
big year has simply been deferred.
Move to green network
For the last couple of years, the industry has been complaining
about sprawling network and server equipment in the datacentre.
With the server world in the middle of a virtualisation feeding
frenzy, it is no wonder that network equipment suppliers want to
get in on the act.
In March, Cisco launched the aggregated services router (ASR)
1000, an edge-class router aimed at both service providers and
large companies, designed to combine different services such as
firewall, virtual private network (VPN) and deep packet inspection
in a single box.
Juniper also announced plans to get into the datacentre with its
edge routing equipment, which is a significant green move because
it simplifies the aggregation layer of a network, says Keith
Humphreys, managing consultant at analyst euroLAN.
"It is a neat model, the way it collapses the traditional
network architecture into a simpler one," he says. "You do not have
so many aggregation points."
ISPs struggled to keep up
Internet service providers (ISPs) and carriers faced business
pressures as people accessed increasing amounts of content. This
year, with the increasing success of sites such as Hulu and tools
such as iPlayer, the problems began to crystallise, warns Sharifa
Amirah, principal consultant at Frost and Sullivan. "2008 is the
year where we finally saw things coming to fruition.
"The plasticity of end-user content has become more prominent in
the business models of not just the telcos but also internet-based
companies," she says, arguing that mobile users in particular are
consuming content in spades, and that carriers she speaks with are
panicking as they try to keep up.
"The big problem is the infrastructure but then there is also
the cost of it. And while consumers are increasingly demanding
high-bandwidth services, there is a low propensity to pay for those
services," she adds.
As consumer appetites for video and other high-bandwidth content
continues to grow, the tensions are likely to worsen.
The internet is broken
The internet may not be perfect, but we try to live with it. But
then, pesky researchers occasionally come along and tell us what we
would rather not know - that the system is inherently broken. Dan
Kaminsky, director of penetration testing at security consulting
company IOActive, knew he was onto something in March when he
discovered a fundamental flaw in the way that the domain name
system (DNS) worked.
The flaw would allow attackers to persuade a recursive DNS
server to refer to their malicious DNS server when performing a DNS
lookup, rather than querying the correct one.
Because this attack could work all the way up to the top level
domain, it meant that in principle, attackers could essentially own
.com or any other domain extension. Kaminsky marshalled together
influential organisations including the United States Computer
Emergency Readiness Team (US-CERT), ISPs, and multiple large web
players to install a rapidly-developed patch that would make it
much more difficult for attackers to successfully compromise the
system.
Days after Kaminsky unveiled full details of the DNS flaw at
Black Hat in August, researchers Anton Kapela and Alex Pilosov
demonstrated another problem at sister conference Defcon.
This time, the bug was in the Border Gateway Protocol (BGP), and
it could be used to eavesdrop on traffic by routing it via an
intermediary server.
The significance of both these flaws is that they are not coding
bugs - they are basic design flaws in the way that DNS and BGP
work.
Fixing them at that level would require going back and
recrafting the initial specifications, and with the whole internet
already running on them, that is very difficult to do.