I remember writing a column for a long since deceased IT publication, where I was discussing the rebirth of the mainframe as a network server, I remember it, not simply for the content, but more specifically the context, and how the sub-editors, in the name of formatting and pure ignorance, change my column title: “The mainframe is dead, long live the mainframe” (quite a witty variation I though!) to simply: “The mainframe is dead”. Not quite the same meaning…
Fortunately, no one edits these blogs (normally!) so hopefully the title here stays as originally typed. The point is, the “mainframe” is being reinvented again, but this time it’s not the mainframe itself that is regenerating itself, but the “network”, so the exact inverse of what happened before.
This thought was reiterated in recent conversations with the excellent Danny Yeowell of Dimension Data (a man who can describe an incredibly complex company structure in simple layman’s terms in the space of 90 seconds!) and in recent work I’ve started with Cirba, a company focusing on “software-defined infrastructure control solutions” AKA managing and optimising virtual storage through software. The point is, whatever a vendor means by SDN and NFV, what is happening is that network functionality is being distributed across the ether, as one giant set of components, manageable from a single, remote entity, whose applications are largely accessed through a browser, regardless of the access device – AKA the rebirth of the mainframe as an organic network, sitting as a software layer, controlling application and data access, wherever the apps and data reside.
As a fundamental consequence of this, storage and networking are becoming ever more integrated, hence the number of acquisitions in recent years of network technology by storage vendors and vice-versa. Looking at Cirba’s solution, while focused on virtualised storage, networking elements such as workload routing, load-balancing and interfacing with NFV elements are all fundamental parts of the product. Cirba’s analytics automate VM routing decisions based on all the required constraints including workload utilisation, business, technical, software licensing, and complex storage requirements.
A side-effect of the likes of Cirba is that it means that the storage vendors don’t have to do this stuff for themselves; they can simply integrate with a “Cirba” as exemplified by this week’s announcement that Cirba is now integrated with NetApp‘s OnCommand Insight OCI), thus providing the aforementioned optimisation to NetApp customers. So the storage companies increasingly become part of the SDN/NFV movement – there is no escape!
Lest we forget to bring “cloud” into this blog, Cirba also provides cloud infrastructure management teams with visibility into when resource shortfalls might adversely affect associated VMs and where excess resources exist for – in this case – NetApp and other storage infrastructure connected to NetApp OCI.
On a more generic level, all this is being put to the test currently by yours truly, so look out for a report on the topic in the near future.