We woke up very early today to attend a packed morning full of networking bonanza, especially Wi-Fi. Since this is not much my activity field, I will have only short entries for these sessions. However, I went around the exhibitor booths to visit SimpliVity, SolidFire, EMC and finally my good friend Francesco Bonetti gave me a detailed overview of PernixData Architect.
3800/2800 Series Cisco Wireless APs
Brian Levin
We were handed over the latest 2800 and 3800 Series APs. Possibility to connect 2 sets of antennae. In high ceiling environments there will usually be one set of antennae. Flexible radio assignment architecture.
3800 AP
802.11ac Wave 2 – nearly internet level speed. Maximized throughput through splitting the bandwidth to several devices and speak to them simultaneously. Allows 5.2 Gb/s per AP (faster than copper 1 Gb/s!). Countries are adding more channels to offer more potential throughput. @networkautobahn had a clever question about how the 5.2 Gb/s is sent back to the network infrastructure. The device has two RJ45 ports and is 5 Gb/s capable (NBase-T).
While it’s not my field, I found the presentation interesting. After all, we’re all Wi-Fi users 🙂
CMX Cloud
I found the presentation very abstract for me. I understood that this is a way to manage existing APs through a cloud-based solution but wasn’t following much the advantage or why it should be exhilarating. Then again, I’m not a wireless guy! Colleagues Peter-Paul Engelen and Dominik Pickardt where much more active than I was.
Cisco ACI Micro Segmentation
Juan Lage, Principal Engineer
This was an interesting session, where Juan explained how Micro Segmentation with Cisco ACI works. These were my three takeaways:
- Micro Segmentation is about defining security in segments that extend beyond VLANs and subnets (applications, application groups, VMs…)
- Have the granularity to define policy between these segments
- Automated, Independent of location
vSphere, Hyper-V and KVM are supported but for the latter two, not all the features are baked in yet. Also, currently policies (or contracts to use ACI terminology) apply to VMs, in the future this could be extended to VMware clusters (or any other hypervisor cluster).
Threat Defense Techniques
Here again not my field of activity and nothing much I can meaningfully contribute about except for one fact. The presenter had well defined use cases, she whiteboarded a lot, and knew very well her product.
Big Data
We were also served a presentation on big data which recapitulated the drivers for big data and how Cisco portfolio can cover these needs but there was nothing new I could’ve been craving for.
While the formal Tech Field Day Extra agenda ended with the Big Data presentation, we kept going along for lunch then delegates went around either visiting booths, discussing, participating etc. For my part, I went to visit the following vendor booths:
SimpliVity
After yesterday’s discussion I engaged with one of SimpliVity SE’s to see their technology in action in a demo lab. I was impressed by the very fast way their solution can create backups/restores even across datacenters. Same for cloning a VM. Powerful tech. Of course with every tech there are prerequisites or assumptions to be taken in consideration, and also some caveats but overall I liked what these guys do. They deserve a better PR. They also need to educate their customers that yes they’re hyper-converged, but that they’re also a storage vendor. The very concept of “convergence” or “hyper-convergence” implies that two things or more are converging to become one, in our case compute and storage.
SolidFire
This was my first intro to SolidFire after yesterday’s informal talk with Jeremiah Dooley. This product is interesting not only for providers (their original market) but also for enterprises where there is a need for IOPS/throughput throttling and segmentation. This can make a great use case in environments that are all about chargeback. It can also in certain cases help isolate some offending workloads and confine them within a set of given IOPS boundaries. This proactive throttling is much more efficient than SIOC (Storage IO Control) that is in fact a contention mechanism. SIOC kicks in when contention is observed and at this point traffic will be prioritized according to “shares” that were assigned beforehand. The built-in throttling in SolidFire is one of the very nice things I’ve seen during Cisco Live Europe. Lovely. And it seems that NetApp really love SolidFire after their acquisition, which is a good signal.
VxRail
We had an overview of the latest VxRail solution (with this Dell-EMC-VCE-VMware thingie I’m getting confused in who is marketing what). Neat hardware and good looking configuration wizard. I didn’t get to see much of the management interface, alas.
My opinion: this product could have been a game changer 18 months ago but now it’s “just” a newcomer to a market dominated by Nutanix, SimpliVity and the likes. Certainly loyal VMware/EMC customers will buy into it, and it is also a technically capable solution, but I am really doubtful on what advantages and what sort of revolution the introduction of this product line is supposed to cause. I would rather applaud VCE/EMC for facing the reality (i.e. that hyper-converged is not a fad but a real market segment) and for having the courage to get over years of selling arrays & taking the plunge into software-defined storage (if one may call it like this – excuse my noobness).
PernixData Architect
While discussing about challenges with data latency in brownfield deployments, Francesco shared with me his experience with PernixData solutions (FVP and Architect). The outcome was a very rich discussion where he explained the challenges solved by these. The diagram below highlights the normal data path of a VM’s I/O from its generation to the moment a write is committed to a disk.
FVP
PernixData FVP eliminates latency and bottlenecks by leveraging either SSD or RAM at the hypervisor level. Writes will be directed to the SSD/RAM, sent to a peer (another ESXi node with FVP running, for high availability), then the write will be acknowledged and info will be sent back to the VM, while in the background the write is transferred to the storage array. In case of read I/O, the data that is read will be cached on SSD/RAM and any further reads will be served from these instead of the storage array.
Architect
Per Francesco’s explanation, Architect is a Big Data Analytics solution built on top of PernixData FVP. Architect draws on FVP to provide recommendations, determine the I/O profile of accelerated VMs, to help troubleshooting and to assist in better design/redesign. I will need to have a look into both products and see how they can potentially be leveraged in my environment.
That is all for today. I may review this post later on to review, edit, amend. Big thanks go to Stephen Foskett and Tom Hollingsworth for making this possible. Overall, the Tech Field Day experience was incredibly stimulating for me as you’re brought to deal with things you may not necessarily know and you get to ask either candid (or sometimes plain stupid) questions. The interactions with people from a different technical background also helps put things in perspective, exchange and enrich mutually.
Disclosure: I was invited at Cisco Live Europe and Tech Field Day Extra by Gestalt IT. Gestalt IT has covered travel, accommodation and food & provided me with a Cisco Live Pass. I have received no compensation for participation in this event, and I am also not obliged to blog or produce any kind of content. Any tweets, blog articles or any other form of content I may produce are the exclusive product of my interest in technology and my will to share information with my peers. I will commit to share only my own point of view and analysis of the products and technologies I will be seeing/listening about during this event.