2. In 1960 to 1985 we started with mainframes and supported a customer base of
about one million users. In 1985 to 2009 we moved to the personal computer ,
client / server model and the LAN /Internet model supported a customer base
of hundreds of millions. In 2009 to 2020+ the industry has completely
changed. We have a variety of platforms ( Mobile, Social, Big Data & Cloud )
with billions of users and it is estimated that the new I.T industry will be worth
£4.8T
The customer has changed and is making us change our networks. Content is
doubling over the next two-year period and emerging markets may overtake
mature markets and we expect 5,200 GB data / person created in 2020. These
new demands and trends are putting a lot of duress to the amount of content
that is going to be created and how we serve and control this content poses
new challenges to data networks.
3. Big Data, Applications, Social Media, and Mobility are forcing architects to
rethink the way we engineer a network. Now we should concentrate more
on scale, agility, analytic and management.
Data Center design was based on the 80/20 traffic pattern rule with
Spanning Tree Protocol ( 802.1D ) where we have a root and all bridges
build a loop free path to that root resulting in half ports forwarding and
half in blocking state. Completely wasting your bandwidth. Even though
we have the ability to load balance based on a certain number of VLAN’s
configured and forwarding on one uplink and another set of
VLAN’s forwarding on the secondary uplink we are still faced with the
problems and scalability of having large Layer 2 domains in your data
center design. Spanning tree is not a routing protocol, it’s a loop
prevention protocol and as it has many disastrous consequences it should
be limited to small segments of the data center.
4. The traffic patterns have shifted and the architecture needs to
adapt. Before we focused on 80% leaving the DC while now a lot of
traffic is east to west and staying within the DC. The original traffic
pattern made us design a typical style of a data center with access, core,
distribution based on Layer 2 leading to Layer 3 transport. The “route
when you can” approach was adopted as Layer 3 aka routing adds
stability to the Layer 2 by controlling broadcast domains and flooding
domains. The most popular data architecture in deployment today is
based on very different requirement and the business is looking for
large Layer 2 domains to support functions such as v-motion.
We need to meet the challenge of the future applications and as new
apps come out with new requirements, it isn’t easy to make adequate
changes to the network due to the protocol stack used.
5. The problem is that we rely on spanning tree, it was useful before but its
past its date. The original author of spanning tree is the now the author of
THRILL ( replacement to STP ).
STP ( Spanning Tree Protocol ) was never a routing protocol to determine
the best path, it was used to provide a loop free path. STP is also a fail
open protocol ( as apposed to a Layer 3 protocol that fails closed ) and one
of spanning-tree’s biggest weaknesses is that it fails open i.e. if we don’t
receive a BPDU ( Bridge Protocol Data Unit ), we assume we aren’t
connected to a switch and we start forwarding on that port. When we
combine a fail open paradigm to a flooding paradigm that can be
disastrous.
6. To overcome the limitation of some are now trying to route ( Layer 3 ) the
entire way to the access layer which has its problems too as there are some
applications that require L2 to function e.g clustering, stateful
devices. However people still like Layer 3 as we have the stability around
routing. You have a true path based routing protocol managing the network
not a loop free protocol like STP and routing also doesn’t fail open and
prevents loops with the TTL ( Time to Live ) fields in the headers.
Convergence around a failure is quick with improved stability. We also have
ECMP ( Equal Cost Multi Path) paths to help with scaling which translate to
scale out topologies. This allows growth of the network at a lower cost. Scale
out is better than scale up.
Whether we have a small or large network , having a routed network over a
Layer 2 network has its clear advantages.
The way we interface with the network is also cumbersome and it is
estimated that 70% of failures on the network are due to human errors. The
risk in changes to production network leads to cautious changes which can
slow down processes to a crawl.
7. STP based Layer 2 has stability challenges; it fails open. Traditional
bridging is controlled flooding, not controlled forwarding so it
shouldn’t be considered as stable as a routing protocol. Some
applications require Layer 2 but people still prefer to use Layer 3. The
network infrastructure needs to be flexible enough to adapt to: new
applications / services, legacy applications / services, organisation
structures. There is NEVER enough bandwidth and we cannot predict
the future application driven requirements so a better solution would
be to have a flexible network infrastructure. The consequences of
inflexibility slow down deployment of new services and applications and
it restricts innovation. The infrastructure needs to be flexible to the
applications. Not the other way around. It also needs to be agile
enough to no longer be a bottleneck or barrier to deployment and
innovation.
8. WHAT ARE THE NEW OPTIONS MOVING FORWARD?
We have Layer 2 fabrics ( Open standard – THRILL ) which are changing the
way the network works and enables a large routed Layer 2 network. A
Layer 2 Fabric for example Fabric Path is layer 2 but it acts more that of
Layer 3 as it’s a routing protocol managed topology. As a result there is
improved stability and faster convergence. It can also support massive ( up
to 32 load-balanced forwarding paths versus a single forwarding path with
Spanning Tree ) and scale-out capabilities.
If we already have a Layer 3 core and we need to support Layer 2 end to
end we could go for an Encapsulated Overlay ( VXLAN, NVGRE and STT). we
have the stability of a Layer 3 core and the familiarity of but can service
Layer 2 end to end using UDP port numbers as network
entropy. Depending on the design option used it basically builds a L2
tunnel over a L3 core. A use case for this would be if we have two devices
that need to exchange state at L2 or if we require v-motion. VM’s cannot
live migrate across L3 as they need to stay in the same VLAN to keep the
TCP sessions intact.
9. Software Defined Networking is changing the way we interact with the
network. It provides us with faster deployment and improved control. It
changes the way we interact with the network and we now have more direct
application and service integration. Having a centralized controller we can
view this as a policy focused network.
Many big vendors will push within the framework of converged
infrastructure ( server, storage, networking, centralized management ) all
from one vendor and closely linking hardware and software ( HP, Dell, ,
Oracle ). While other vendors will offer a software-defined data center, in
which physical hardware is virtual, centrally managed and treated as
abstraction resource pools that can be dynamically provision and configured
( Microsoft ).