About Rowell Dionicio

Wi-Fi expert. Coffee addict ☕️. Tech nerd.

UniFi Adoption

November 13, 2018 By Rowell Leave a Comment

There is a strong demand for affordable enterprise-grade wireless solutions. Ubiquiti UniFi appears to be filling that need.

Ubiquiti may be well known for their point-to-point products but recently, there’s been a decline in that product line and a gigantic increase in the UniFi solution. UniFi rivals the larger enterprise wireless providers but offers it at a reasonable price.

A lower price point is enticing enough for many customers. UniFi access points are in the sub-$1k range which doesn’t break the budget.

Recently, Ubiquiti’s Q1 2019 earnings report was released and they saw an increase of 41.3% in revenue for the Enterprise technology segment. That brought the revenue up for that segment to $177.9 million. This includes UniFi, mFi, and AmpliFi products. This is a 15.1% increase in growth from Q1 2018! In Q3 2018, they grew 31.6%, to $149.5 million which was 60% of total sales!

The growth is explosive.

There’s no doubt, the UniFi management dashboard provides more than just the basics of wireless management. Ubiquiti is expanding it to become a true provider for high performing wireless networks.

UniFi’s Impact

I’m worried about poor wireless design. While UniFi may be affordable and easy to configure, it is equally easy to implement wireless incorrectly. High capacity and challenging environments need a proper wireless design to get the performance out of the investment.

The other concern is quality. I’d like to test quality of the access points for myself. Stress tests with competitive access points should also be performed. If the performance is along the same lines as other vendors then we can say UniFi is making a dent in an industry largely dominated by big vendors.

The Shift

Will we see a bigger shift in the wireless industry towards lower cost access points? Yes. I haven’t been impressed in code and hardware quality from the incumbent players. There’s plenty of room for someone to come along and cause disruption. Vendors must provide value beyond simple configuration. Attention must be spent on code quality and hardware to make a clear differentiation.

DeMISTifying Wi-Fi Issues

September 23, 2018 By Rowell Leave a Comment

Mist Systems first delivered a new way of looking at Wi-Fi at MFD2. I was impressed with their product and I continue to be impressed with the progress they’ve made since their last time at MFD.

What Mist Systems brings to the table is a way of managing Wi-Fi the way it should be in our current state of the industry and workplace. The amount of confidence in their product shows. They’re truly proud of it.

I’ve been confused about how vendors use Artificial Intelligence (AI), and I’ve never really known if they really were, but Mist definitely goes into it with their platform. And the expandability of the Mist product, the amount of data collected, and how that data can be used via the API is astounding. They give you full flexibility and autonomy of how you want to handle that data.

Confidence

Sudheer brings a lot of energy on stage. He knows the product very well and can answer any question. As he should. He exudes his energy and it feeds the room. The confidence really shines and helps sell Mist and the problems they solve. Sudheer is not afraid of saying how they’ve replaced incumbent competitors in existing environments.

Mist brings a lot of issues to light. Issues that were not seen by Meraki or Ruckus and where bubbled up as Wi-Fi issues.

One such customer they are proud of is Pet Smart. They went head-to-head with Meraki. Sudheer described Pet Smart’s attitude as skeptical as they entered into the bake off. Now they are rolling out Mist across all their stores, identifying and resolving issues quickly, and leveraging the Mist API to provide better customer experience.

Is it believable to say that Mist can really achieve all those check boxes with just a Wi-Fi solution?

With many large companies moving to Mist, there seems to be a similar story amongst all of them. These companies don’t know what’s happening on their networks. And Mist has provided an answer to their pain point that appears to be a Wi-Fi issue.

There’s another conversation that extends from that, which is if these networks were properly designed and the engineers had troubleshooting expertise, would they have needed Mist?

Then there’s the age of Mist themselves. Being less than 5 years old, how can they be trusted? That age can be deceiving. The company may be young but it’s being ran by seasoned professionals. Bob Friday co-founded Airespace and later held roles within Cisco. Sudheer Matta also held roles at Cisco within the wireless business unit.

Platform

The platform itself is another cloud managed Wi-Fi service. They’ve built it around open source technologies which I have no real expertise in critiquing.

But this platform provides a way to bring a constant stream of updates, bug fixes, and new features through their dashboard. From the controller-based Wi-Fi world, a consistent stream of updates and feature enhancements is unheard of.

Mist does this at scale with microservices. Being in the cloud world, microservices may be a simple topic but in the way I understand it this provides a way to role out updates easily and to fail fast and safe without causing global customer disruption.

API

What has me really interested in Mist is their API. Compared to Meraki’s API, you’ve got yourself a full-on toolkit at your disposal. It isn’t baked on top of a product.

The API was developed alongside the product. What’s really surprising to me is some customers opt to build their own dashboards and skip Mist’s altogether.

The Mist API is extensive and can cater to a novice or seasoned professional. If there’s specific data you’re looking to use, manipulate, or analyze, you can do it with the Mist API.

Artificial Intelligence

Mist is going all in with Marvis, their AI-driven Virtual Network Assistant (VNA). Marvis, which draws similar ties to Iron Man’s Jarvis. When it comes to AI, we would need a system that would perform functions that normally require human intelligence and inspection. Marvis is the AI that will help look up issues for you, perform anomaly detection, faster root cause analysis and real-time changes through RRM.

Some of the ways Marvis can be utilized by IT teams is to ask it simple questions such as:

“Unhappy users at site office”
“Troubleshoot site office production”
“List client events yesterday”

There’s so much trust that goes into Marvis that the Mist support team uses Marvis for troubleshooting with any client tickets that come in. They eat their own dog food. As a result, Marvis brings up a lot of data for correlation and provides enough to make it actionable.

Future of Mist

How bright is Mist’s future? What concerns me with being cloud-based, gathering immense amount of data that has to be analyzed for AI and Machine Learning, is how will this amount of data impact Mist financially? Cloud is great for enterprises and we know the benefits. Big Data in the cloud can get expensive. Will we ever see an on-premise solution for even the biggest customers? Who knows.

From what I can gather, Mist is not for those looking for a single vendor stack. Mist doesn’t have a switch, router, or firewall. But what they do have is a way for you to leverage API to work with their partners, Palo Alto Networks and Juniper. Mist doesn’t want to be a jack of all trades, master of none. They stick to what they know and do it well.

Python Day 4: Dictionaries

September 23, 2018 By Rowell Leave a Comment

Dictionaries

Indexes for dictionaries are called keys. Keys are associated with a value called a key-value pair. A dictionary is unordered. Compared to a List, there is no index starting at 0.

Dictionaries are typed with braces {}. To assign a dictionary to the client1 variable:

client1 = {'rssi': -52, 'channel': 36, 'noise': -96, 'tx rate': 600, 'mcs': 9}

The keys within this dictionary are rssi, channel, noise, tx rate, and mcs. The values for those keys are -52, 36, 96, 600, and 9, respectively.

To access a value through a key it is performed by:

client1[‘channel’]
36

Trying to call a key that doesn’t exist in the dictionary will result in a KeyError error message.

client1['bssid']
Traceback (most recent call last):
File "", line 1, in 
KeyError: 'bssid'

With methods, values can be returned from a dictionary. Let’s say I want to print out the key and value of client1. I’ll use a for-loop to iterate through each of the key-value pairs and print them out using the items() method.

for i in client1.items():
print(i)
('rssi', -52)
('channel', 36)
('noise', -96)
('tx rate', 600)
('mcs', 9)

If we just needed to print out the keys we would use the keys() method. Again, I’ll use the for-loop to iterate through the keys and print them out:

for k in client1.keys():
print(k)
rssi
channel
noise
tx rate
mcs

If we wanted just the values we can do the same for-loop but call the values() method:

for v in client1.values():
print(v)
-52
36
-96
600
9

Python Day 3: List Methods

September 11, 2018 By Rowell Leave a Comment

The previous day went over the basics of the Python List. On this day I go a little further into List Methods.

List Methods allow you to work with the items within a List. With methods such as removing and sorting. Useful methods when working with a list at any capacity.

Removing Values

There’s going to be a time where you need to remove a value from a list. It’s possible to do it with the del Statement.

wifi = ['rssi', 'snr', 'spatial streams']
wifi
['rssi', 'snr', 'spatial streams']
del wifi[0]
wifi
['snr', 'spatial streams']

But it is possible to do it with the remove method as well:

wifi = ['rssi', 'snr', 'spatial streams']
wifi.remove('rssi')
wifi
['snr', 'spatial streams']

The difference is the remove method uses parenthesis with the actual name of the item within the list. It’s easier to call out the actual name than the index number within the list.

But what happens if the item isn’t in the list? A ValueError occurs:

wifi.remove('antenna')
Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
ValueError: list.remove(x): x not in list

Sorting Values

Let’s say I have a list of Tweeps and I want to sort it alphabetically.

tweeps = ['françois', 'robert', 'sam', 'keith', 'alex']
tweeps
['françois', 'robert', 'sam', 'keith', 'alex']
tweeps.sort()
tweeps
['alex', 'françois', 'keith', 'robert', 'sam']

Good stuff. It also works with numbers.

ages = [54, 92, 65, 27, 3]
ages.sort()
ages
[3, 27, 54, 65, 92]

What a list is not capable of sorting is a mix of integers and strings. You end up with a TypeError:

mix = ['spam', 'rice', 3, 5, 'red']
mix.sort()
Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
TypeError: '<' not supported between instances of 'int' and 'str'

I’m sure in networking there will be many reasons to use a list. I can just see it happening but I don’t have any specific use case as of yet. It’s still early for me to figure out what I’d be using a list for. Maybe a list of IP addresses or interfaces of a switch or router.

Python Day 2: Lists

September 10, 2018 By Rowell Leave a Comment

Lists look like they’ll be useful when I get to a point where I can write scripts. Python Lists can hold multiple values in an ordered sequence.

Within the list are items which are ordered by an index. The index starts with 0. Each item is separated by a comma and any strings must have quotes around them.

To assign a list to a variable and retrieve an item from the list the syntax looks like:

groceries = ['oranges', 'spinach', 'eggs', 'milk']
groceries[0]
'oranges'

As I said earlier, the first value in the list starts with 0. It’s possible to go in reverse by using negative indexes:

groceries = ['oranges', 'spinach', 'eggs', 'milk']
groceries[-1]
'milk'

What if you wanted to access multiple values within a list? That’s done by slicing. A slice starts with the first integer, a colon, and the second integer is where the slice ends, but not including that item. Take a look:

groceries[0:3]
['oranges', 'spinach', 'eggs']

Modifying items within a list is possible. Let’s say you wanted to change oranges to apples in the groceries list. We need to modify index 0 within the list.

groceries[0]
'oranges'
groceries[0] = 'apples'
groceries[0]
'apples'

Removing items from a Python list is fairly simple as well by using the del statement. We must reference the index of the item we want to remove.:

groceries
['apples', 'spinach', 'eggs', 'milk']
del groceries[0]
groceries
['spinach', 'eggs', 'milk']

The items in the list move up in their index values which means spinach is now at index 0.

Adding items to the Python list is done by appending it using what’s called a method. Methods are a little different when it comes to lists because we call on the method on a value:

groceries
['spinach', 'eggs', 'milk']
groceries.append('yogurt')
groceries
['spinach', 'eggs', 'milk', 'yogurt']

By appending to the list, the new item we are adding gets added to the end of the list.

It is possible to add items to the list at a specific index. The index value must be passed, with the new item, using the insert method:

groceries
['spinach', 'eggs', 'milk', 'yogurt']
groceries.insert(0, 'pears')
groceries
['pears', 'spinach', 'eggs', 'milk', 'yogurt']

There’s quite a lot you can do with lists. They are very flexible and can be modified with ease. That’s all I’ve done with Day 2 of my Python learning exercise. The next Day will include more Python List work.

On to more practicing…