Archive for the ‘Networking Field Day’ Category
I wanted to start off stating Brocade broke one of the biggest barriers with getting involved with SDN and labbing out the technology. Brocade offers a free download of their Vyatta Controller! With this free download you can run a 5x node SDN network for one year, included with 60x days of support! This eliminates a huge obstacle of actually purchasing the software, sure you may still require the hardware but Brocade SDN Solution features support for OpenDaylight/OpenFlow so you do have many different hardware options.
Now, that I got that out of the way my two favorite pieces of the Brocade was 1. The technical overview of the Vyatta controller and it’s architecture, it was great to see how to the services overlay on each other and what makes it tick. Usually when it comes to some type of SDN solution it’s usually presented as some type of application that does magic. In this case however Brocade definitely did their due-diligence to cover how their controller actually functions. The 2nd thing I loved about this presentation was just how frank and up-front the presentation was. My favorite quote of the whole the presentation was “We know how to code, we went to school. We chose not to program we went into networking.” I can’t say how happy I was to hear someone actually say this! However like it was mentioned in the presentation it appears to be a natural evolution of the field.
As the presentation continues, you really get a sense about how far along the Vyatta controller has come along once the conversation steers towards volumetric traffic management. Having the additional and built-in monitoring of the traffic flows with sFlow and OpenFlow addressing a level of application performance management many current-day data centers frankly do not even have in place today just shows how grown up the tool is becoming. This is built upon again with the flexibility to handle elephant flows differently than other typical data flows, if you are not familiar with the term elephant flows these are just traffic flows that transfer a very high amount of traffic (IE: Something like backup traffic). I can’t tell you how many few companies I’ve worked with in the past that have actually taken into account these ‘elephant flows’.
Now, I don’t want to ruin the whole presentation for you, if you have not watched it yet I highly recommend you give it a watch. There also a great slide in there about Ivan! If you think SDN is still a mystery it’s time to get that Vyatta controller downloaded and running! No more excuses!
You can download the Vyatta Controller here.
Brocade’s Networking Field Day #9 videos can be found here:
I first need to give a shout out to @_vCarly and her amazing skills at the white board, I only wish my white boarding design were half clean as that! If @_vCarly were to a host a white boarding session at CLUS this year I would most likely attend!
I suppose we should discuss the actual presentation now! Like probably many of you I have heard an awful lot about Cisco SDN and Controllers (Cisco ACI) however until now it has all been theory and hear-say. I can finally say I’ve seen the GUI and got to hear a deep dive of how this new architecture is supposed to work. Now, that I have seen it I am believer there is a lot of potential here and allow us to re-think the way we consider designing Data Center networks and here is why I think that:
- Building around the application, with this new design model we are almost forcing the application designer & server administrations to tell how their servers/applications work and communicate on the network. I have a spent a lot of time with AppDev’s going over what their application does and what else the App communicates in order to troubleshoot an issue. This approach removes much of the hassle without ever seeing the hassle. Future engineers might not know ‘the struggle’ of discovering an application via packet analysis and bridge calls.
- Deployment Orchestration – by specifically calling out the application and the service chaining from the get go, deploying a new application is much more streamlined and simplified in the long run when internal process are stream-lined. Eliminating the time-consuming need to:
- Login to multiple networking devices to configure SVI’s & Trunks
- Login to multiple firewalls creating the appropriate rules.
- Login to Load balancers and creating devices pools, health monitoring, virtual servers, and so forth.
- Troubleshooting and flow monitoring – Out of the box this box will be watching flow statistics. So tracking traffic statistics can easily pin-point packet loss and drops. Which may mean bad news some of the simpler NMS solutions out there unless they start looking beyond the typical Up/Down & Interface utilization mantra.
- Something that was brought up on Twitter post-event was the support for SNMP, which is a great question since SNMP was never mentioned (in any of the SDN-related presentations for that matter).
Now, with all that said and the fact Cisco ACI builds a Data Center around EPG’s (End Point Groups) I really look forward to this new design architecture. It’s probably going to be a bit more tedious up front since some applications will need to be called out specifically instead of going “O, that is a web server. Place it on this VLAN” but in the long run this will streamline many different parts of data center operations.
Links to the Networking Field Day 9 presentations below:
One thing I really liked about SolarWinds was the fact they were the first vendor to start with a white board discussion. Being a network engineer I am a big fan of drawing boxes and lines on a white board. To me it is a more engaging way to have a presentation compared pre-built PowerPoint slides with fancy sounds and motions.
SolarWinds initially started with a architecture overview of their product, involving each of their modules and how these modules are incorporated into the entire eco-system. If you are not familiar with this I highly recommend starting with that session.
Now to talk about the new announcements SolarWinds and the features that will be included in their newest releases.
1. Wireless Heat Map – I’ll admit (and this won’t be a surprise to any of my usual readers) I have been following this feature on the Thwack Forums through the different beta’s for some time now and it’s really had my attention since the initial announcement. I was very curious to see how this would match against other solutions such as MSE Controllers and other site survey tools. While this feature might not be as fully featured as the previous mentioned solutions, this is great addition to SolarWinds arsenal and will definitely be useful for smaller shops that run WLANs. However, also keep in mind this is the first implementation of the functionality and if Solarwinds has proven anything over the years they take an idea and run with it!
2. QoE – Quality of Experience. This in my opinion is one of the greatest new features SolarWinds has added in a long time. A trend I have been seeing over the years is the fact network monitoring needs to go beyond the typical Up/Down & Interface utilization model especially with the adapting trend of 10G, 40G, & 100G infrastructure (and there are some multi-GB architecture getting thrown around nowadays to). It’s time to start monitoring closer at the application level and that is what SolarWinds is starting to do.
Now there is definitely some thought that needs to go into the setup of this feature SolarWinds has a few articles discussing it here:
Remember one of the biggest factors of Network monitoring is the perspective in which you are monitoring the network! take into account where you deploy these QoE sensors and how you interpret the statistics!
3. LEM – SolarWinds also did a presentation of their Log & Event Viewer, which I found very useful. I’ve known about this product for a while, however I was not quite sure how it goes up against a typical syslog collection. As it turns out the SolarWinds LEM product is in a league of it’s own touting to be a full SIEM product. One thing that makes LEM a little unique is the fact it is a completely separate & isolated product not tied into any other SolarWinds product.
The Networking Field Day 9 SolarWinds presentations can be found here:
Also, I want to throw it out there. SolarWinds has a Demo Website where you can click around and explore many of their different modules. I highly recommend looking around if there is a modules you are curious about. Found here: Oriondemo.Solarwinds.com
Well, Networking Field Day 9 has ended, it was a very quick, fun and informative three days.
I have to start by giving NetBeez a lot of respect for the great presentation, and I that same I almost want to apologize for the suggestions and feedback that we (All the delegates) bombarded them with during their presentation. They were certainly taking notes of our suggestions and even questioned our suggestions.
Now, NetBeez’s big announcement during this event was the availability of a ‘Wireless’ NetBeez. (If you are not familiar with NetBeez in general I recommend swinging back around to my previous post about their product). Which in my own opinion is a great addition their product. Having your own client node monitoring various key points of the network recording WLAN statistics provides a whole new level of WLAN monitoring that we do not currently have from the client level (or something I have not seen next).
- When we start talking about monitoring WLAN’s I see a lot of untapped potential to spearhead into many different industries:
- Warehouse / Inventory
These industries usually consider their WLAN’s and their accompanying WLAN Applications to be mission critical, and without these key elements the entire operation can shut down costing large amounts of money. So I can easily see NetBeez pushing into these industries. One thing I want to see is a more rugged-ized model since these environments don’t tend to be very inviting.
Being able to track both application & WLAN level statistics from different locations of the wireless network can identify issues before they happen either at the access point or from an upstream distribution switch. Which to me offers a great value add to anyone who must manage large distributed WLAN’s. This little device provides some much needed visibility across the entire network path incorporating but Wired and Wireless statistics.
I am looking forward to kicking the tires on this NetBeez product and I am looking forward to any new features that gets rolled into their products.
If you missed the live presentation, you can view the recordings here:
If there is one topic prevalent at Networking Field Day #9 it is SDN, well okay maybe this is one of the most prevalent topics in the networking industry right now. Among many the presenters at Networking Field Day #9 quite a few of them bring some flavor of SDN or SDN-WAN’s to the table.
A few of the presenters:
- Cumulus Networks – A Linux based SDN-powered data center we definitely have a great foundation here. While looking around Cumulus’ website I really do like the sheer amount of information available on their website.
- Validated Design Guides – Pretty extensive design guides covering guidelines for implementing Cumulus network devices with VMWare, OpenStack, & Big Data (Which are definitely worth reading)
- Looks like they recently released a new version v2.5, and they have presentation available here. Definitely some good information in there.
- Probably the most important option (in my option), is the ability to ‘test drive’ this solution. You can request access to their remote lab to get a better look and feel for the Cumulus Solution.
- Considering this a *nux Distro, some of the backend process are common in the industry. Namely Zebra & Quagga for instance. It’s been a while since I’ve worked these, so whether we consider SDN a brand new technology or a shiny new tool to configure your network. Deep down it still utilizes the same backend methods you and I are familiar with.
- CloudGenix – An SDN-WAN Solution, this is an interesting twist of the SDN perspective because up until now when I heard the term SDN I usually associate it within the Data Center.
- A few noticeables, you can request beta access here, and stay up to date with CloudGenix’s progress.
- I really like how companies are embracing this openness.
- To me this sounds like it could be a very useful & popular take on the SDN-perspective, there are many companies out there with 100’s of remote sites that could definitely benefit from the SDN model. Keeping policies in sync across that many locations is always a tough business challenge and drastically affects the troubleshooting process, so I can’t wait to see CloudGenix’s approach and how they can aid in this issue.
- Using the SDN with the WAN in my opinion may also present a different set of challenges.
- Where as you typically have a stable and reliable underlay within your data center, the enterprise WAN may or may not have that same level of reliability & stability in the WAN underlay. Does that introduce any new issues to the SDN model, and if so how does CloudGenix approach those new challenges?
- A few noticeables, you can request beta access here, and stay up to date with CloudGenix’s progress.
Both of these companies should be presenting unique and interesting presentations, after all SDN certainly has a lot of buzz in the networking industry and now we are looking at different flavors of SDN. These different flavors of SDN could really affect the future of network implementations. Now a small part of me is wondering, can I get an SDN-Integration application to make sure my SDN-WAN and SDN-Data Center mingle in harmony? Cause we all know sometimes what is good for the goose is not always good for the gander.
With Networking Field Day #9 quickly approaching, I can’t help but start digging into the presenters and see what they currently offer.
In reviewing the presenters at Networking Field Day #9, one of them NetBeez offers an interesting take on network monitoring. I must admit this is one of the presenters I needed to do a little research on. Now, when I first went to the NetBeez homepage, the first thing I noticed was what they provide and bring to the table:
How many companies state they provide ‘end-user monitoring’ this to me is a unique twist on network monitoring. Think about it we have Application Performance monitoring, network device monitoring, traffic flow monitoring, & many others but end-user monitoring is not usually something you hear about too often.
Like many of us have, I have been using IP SLA across many networks to give me a perspective as to what the user what might be experiencing. While this is a good idea it is however slightly flawed in my opinion, my reasons below:
- It’s not really from the user’s perspective. (Usually)
- Usually the IP SLA is configured on a distribution or core device. NOT from the access layer.
- IP SLA does not scale very well.
- How are you going to manage all these IP SLA’s?
- You need to be careful when/where configuring or terminating a large number of IP SLA’s on individual routers, they require resources to run! You can very easily go overboard!
- Reporting and monitoring.
- Can you report on the IP SLA statistics? How long were response times above a certain threshold vs. within a reasonable expectation?
However, after reviewing NetBeez I see some potential. As their solution does appear to address some of those concerns & flaws:
- True user perspective
- NetBeez devices can be connected directly to the access layers, providing a network perspective AS A REAL USER. Albeit at the cost of a single switchport.
- The NetBeez devices are managed from one central management console, easing the management of this monitoring solution to a single console.
- Now your infrastructure devices simply need to pass packets in the data plane and service clients not worrying about the additional load of the IP SLA processes.
Now these are just some of my initial impressions, I am quite curious to see this product in action, to see if they live up to my impressions, and what they have in store for the future. In my opinion this product has a lot of untapped potential.
A few things I am curious about:
- Cost – I’ve worked with many companies were I could easily deploy 50 to a few 100 of these little devices. Many of know the budget is a big decision maker, can these useful little guys break the budget?
- Consolidated Data views – Deploying a lot of these guys might get overwhelming quickly and I am curious how efficiently this data can be summarized.
- Reporting – With all this useful information I am wondering if the data is kept for historical purposes and can be reported on later. I think being able to report on response times for the various applications (HTTP, DNS, etc) can easily provide trending capabilities after a 3 month time frame.
- Support for additional applications, or advanced monitoring capabilities- Many companies will be running some type of home grown application, can NetBeez devices be configured to perform simple TCP handshakes over specific TCP ports? What about taking HTTP monitoring a step forward and navigate down a few pages instead of a single HTTP GET, maybe even reporting on individual page elements. eCommerce is a very large and metric driven business that could definitely benefit from advanced HTTP monitoring.
- Most importantly can we run monitors between different NetBeez devices
Now, don’t think of those curiosities too harshly, network monitoring is a topic I tend to be overly critical about (Just ask anyone I have worked with)
With all that said I this should be a very interesting presentation, and one I am looking forward to.