The popularity of cloud computing and cloud services has been driving the demand for higher bandwidth, faster speed and lower latency performance. It further pushes the migration of data center switch market to 400G/800G. Read on to know how the data center landscape is evolving, and what it means for data center switches.
The history of data center evolution has shown that there is an ever-increasing change in bandwidth, fiber density and lane speeds. This also has a far-reaching impact on the DC landscape. The transition to 400G/800G is way closer than you think.
Data has become increasingly crucial for individual consumers, organizations, and large economies alike. It penetrates everywhere when we navigate the internet world and connect with others. As data centers are evolving, data usage around the world is shifting too.
The following list shows some major trends in global data usage.
Social media enjoys explosive traffic growth.
Massive small cell densification enables the rollout of various 5G services.
Developments of IoT and IIoT (Industrial IoT) are gaining momentum.
Traditional office-based work is shifting to remote options.
Every data center looks to achieve high-bandwidth, high-availability, and low-latency so as to meet the rising demands of end users. In response, data center architectures are rapidly moving toward spine-leaf and fiber-dense mesh networks, which can help simplify the data center network and deliver higher capacity. More flexible connectivity, higher fiber-count backbone cables and improved scalability enable data center networks to support ever-higher lane speeds. This helps data center operators be more prepared to realize the migration to 400G/800G.
The impact of hyperscale and edge data centers on the overall data center landscape is significant. As demand for lower latency performance increases, hyperscale providers work to deliver cloud computing services closer to the end user and devices.
In edge data centers, the shorter distance between a user and the processing network allows data to travel faster, ensuring instantaneous and optimum access to data and applications with lower latency. At the same time, growth in hyperscale and edge data centers pushes smaller data center owners and operators to adapt their infrastructures to support the ever-increasing traffic demands.
For data centers, performance is actually a matter of checks and balances among servers, switches and connectivity. The evolving data center landscape certainly has a direct impact on data center switches. Network operators are all looking for highly scalable, reliable, and high performing switching solutions.
One crucial change is the move from a top-of-rack (ToR) architecture to either middle-of-row (MoR) or end-of-row (EoR) configuration. Compared to ToR architecture, MoR and EoR need fewer switches deployed in racks. It means less rack space, less maintenance effort, hence less power consumption and cooling systems.
In addition, migrating to higher speeds involves simply replacing the server patch cables instead of replacing the longer switch-to-switch links. However, the shift requires a structured cabling approach that helps facilitate the connections between the in-row servers and the MoR/EoR switches.
Closely related to the shift from ToR to MoR/EoR is the increase in switching speeds. Serializers/deserializers (SerDes) that provide the electrical I/O for the switch application-specific integrated circuits (ASICs) are expected to hit 100G once IEEE802.3ck becomes a ratified standard. It means switch ASICs are also increasing the I/O port density.
Higher switch-port densities support more network device connections, offering the potential to reduce top-of-rack (ToR) switches. This, in turn, reduces the overall number of switches needed for the data center architecture.
More and more data centers are beginning to adopt 200G/400G switches, QSFP-DD transceiver modules, and 400G fiber optic cables. This year Google is expected to debut 800 Gb/s switch shipments, driven by the availability of 800 Gb/s optics.
In 2021, the data center switch market unfolded pretty much in line with Dell’Oro Group's expectations. Data center switch sales across the cloud segment grew by double digits, while the non-cloud segment increased in the mid-single digits. The global data center switch market is expected to reach USD 19.9 Bn by 2026, growing at a CAGR of 5.6%.The market appears to have proven resilient to the effects of the COVID pandemic.
Dell’Oro Group has projected that the data center switch market will grow by double digits in 2022, with the cloud segment growing almost at twice the rate of the non-cloud. Major drivers for such a robust market forecast include:
Panic purchasing behavior fueled by ongoing supply challenges
An accelerated pace of digital transformation
Expansion cycles fueled by new AI workloads
The rosy picture for data center switch market certainly brings along great opportunities, but there are also challenges lying ahead in the transition to 400G/800G.
More pronounced silicon diversity
Silicon diversity has been a theme in data center industry over the past few years. It is fueled by the need to put pressure on Broadcom, a semiconductor giant that dominates chips for networking equipment to date.
It is expected that an increased number of viable merchant silicon suppliers such as Cisco and Marvell/Innovium, coupled with industry-wide supply constraints, will further accelerate silicon diversity in 2022. At OCP 2021, Cisco announced that it will be supplying Meta (formerly known as Facebook) with its Silicon One chips on the Wedge400C for Top of Rack applications.
Surges in Demand for Smart Devices
Mobile networks are used to connect all sorts of devices such as smartphones, smart TV, cars, home appliances. Technological advancements in those smart devices are driving the demand for sophisticated connectivity and enhanced networking solutions. Therefore, there is a rise in the adoption of IoT & cloud computing. The surge in demand for smart devices is expected to drive the integration of the chips in the data center servers, which will provide lucrative growth opportunities for the global data center switch market.
High data center operational cost
One of the challenges that hamper the growth of the data center switch market is the high operational cost of data centers. As energy expenses account for a big share of the operational cost, data centers located in various parts of the world need to consider the energy prices in the area. For hyperscale data centers and cloud service providers, the energy expense alone can be exorbitant. In addition, other operational costs such as machines, maintenance, labor also restrain the market growth.
Complex architecture and compatibility issues
Sever virtualization, computing, storage, and cloud technologies have been deployed to help improve scalability and flexibility in the data center. This inevitably makes the data center architecture more complex. Although data center switches with high performance and increased bandwidths are capable of handling large workloads, there is still much complexity involved in implementing high-bandwidth solutions across different architectures.
Furthermore, it becomes challenging to create compatibility between different data center technologies, which may incur substantial costs and delay in new deployments.
It's likely that the data center switch market spotlight will continue to shine. For data center owners and operators who are planning for 400G/800G migration, they need to take a deeper dive into where things stand now and keep track of how things are evolving.