Ask the Expert

SAN vs. LAN network integrity

I have read that the Ethernet network on which NAS is based does not track the request sent to the server and acknowledge that it was received. In the case of the SAN, the writing guarantee is given by the switch. Is that right? Why and how?

    Requires Free Membership to View

    Login

You bring up a good question about SAN vs. LAN network integrity. There has been a lot written about the pros and cons of storage-based networks like Fibre Channel and IP-based networks, including Ethernet, from pundits on both sides of the fence. In general, storage-based networks including Fibre Channel have been designed with deterministic performance behavior and data integrity as primary requirements, while IP-based networks have been designed with good performance, low cost and standardization for mass adoption as primary criteria.

Storage-based networks are designed from an undersubscribed approach, meaning that, if anything, performance and data integrity are over-engineered, which has a higher cost. What this means is that it is possible for an IP-based network to encounter network congestion resulting in a dropped packet, thus forcing data to be retransmitted, impacting performance. Given the wide spread deployment and adoption of IP-based networks supporting storage-over-IP for mission-critical applications, including NAS, iSCSI, iFCP and FCIP, vendors have taken precautions to insure data integrity and protection should a packet drop occur. To learn more about storage and network interfaces, protocols and access methods check out chapter 3 ("Networking with your storage") and chapter 4 ("Storage and I/O networks") in my book Resilient Storage Networks.

This was first published in December 2004

Join the conversationComment

Share
Comments

    Results

    Contribute to the conversation

    All fields are required. Comments will appear at the bottom of the article.
    Back to top