The rise of IP storage

Enterprises are demanding more flexible, open network architectures as they adjust to the ‘3rd Platform’ era to meet the demands of mobility, cloud services, big data analytics and social networking.

According to IDC, to meet these new requirements, organisations should consider dedicated networks for IP-based storage as a better approach to align their networking infrastructures with business needs.

Old assumptions about how to design a network that connects application hosts to their network-attached storage arrays are subsequently changing.

The scale of IP storage, both block and file, has greatly expanded as more applications than ever before are using files for managing storage. At the same time, ethernet bandwidth has increased to meet the demands of server virtualisation, hosting multiple applications per physical server.

Today, many businesses have as much or more NAS storage as they have block storage connected to their applications. And a growing number of applications that are using IP storage are mission critical.

The road to IP storage networks

The introduction of an open protocol for storage input and output (I/O) channels, Small Computer System Interface (SCSI), has now become ubiquitous. However, as storage volumes grow, this architecture cannot keep up with requirements.

SCSI’s need for low deterministic latency and guaranteed delivery has contributed to the creation of two other solutions.

One is Fibre Channel Storage Area Network (FC SAN), which acts as a lossless, low-latency network for SCSI traffic. Another solution is to use shared file systems such as Network File System (NFS) and Common Internet File System (CIFS). These solutions have traditionally used a ‘best effort’ IP network with Transmission Control Protocol (TCP) to provide connectivity and ensure delivery of the data.

The evolution of dedicated block I/O channels from point-to-point busses into any-to-any networks has met the need for dramatic growth of data storage and larger bandwidth.

Why use a dedicated IP storage network?

As IP storage capacity is nearly doubling every two years and increasingly being used for more mission-critical applications, organisations should consider dedicated networks for their IP-based storage as a better approach to align their networking infrastructures with business needs.

Co-mingling storage with other network traffic results in network configuration compromises that can complicate storage management and increase corporate risk, which is why it is recommended that IP storage traffic use a dedicated physical network.

A prime example of the need to separate out IP storage traffic is that application performance can be gated by many things; but slow response to a read or write of data in storage is devastating to the network.

For IP networks, little variation on the network is highly desirable to ensure predictable application performance as the workload varies. Other reasons include:

• low, deterministic latency;
• guaranteed delivery;
• a smaller administrative domain, which is easier to troubleshoot;
• fewer configuration compromises;
• better fault isolation; and
• less complexity to upgrade and maintain.

These reasons can be easily backed up when problems arise. When diagnosing technical problems, the smaller the domain of analysis, the easier it is to perform the analysis.

This supports the idea of a dedicated network for storage traffic.

Upgrading and maintaining a large network can also become complicated. It is easier to maintain smaller scale networks that are dedicated to storage traffic than to maintain a large, complex network that handles co-mingled storage and general-purpose network traffic.

For larger environments, there are compelling advantages to having a separate physical network for IP storage traffic, regardless of whether you use block storage (iSCSI) or NAS storage (CIFS and NFS).

Congestion on the network can result in a ‘slow start’ as the application tries to determine the network performance, which can take seconds before full flow is re-established after a congestion event which manifests itself as slow storage access or application performance degradation.

Low and deterministic latency is dramatically affected by small-duration traffic spikes, resulting in dramatic decreases in the transaction rate for applications as they wait for storage I/O to complete.

TCP is required to achieve guaranteed delivery, but the side effects of the slow start mechanism can cause orders-of-magnitude increases in I/O latency for minutes at a time, dramatically slowing application response.

The process of diagnosing and eliminating this behavior can be very complicated and time consuming. Therefore, simplifying the data centre network design by using a dedicated IP storage network for both block and NAS storage pays dividends.

Conclusion

Today’s IP storage applications are becoming increasingly complex, business critical and high profile. Failure is not an option, but running these bandwidth-intensive, latency-sensitive applications on a shared network puts storage performance and availability at risk.

The best practice solution for storage (FC or IP NAS) is a dedicated, fabric-based network, both inside and between data centres. This is essential for maintaining reliable and predictable application performance, as well as providing security for data flows for storage between data centres.

Storage applications thrive on dedicated fabrics that are reliable, high performing, self-healing and easy to manage.