jump to navigation

Ramblings on the Programmable World June 6, 2013

Posted by Edwin Ritter in Cloud Computing, Trends.
Tags: , , , ,

I am an active reader of Wired and enjoy articles that deal with emerging trends. From a recent issue, I found an a write up that relates to big data. In the near future, data will be generated by every day items. The main concept of the piece deals with connecting our analog devices such as a refrigerator (and, everything else) to a network. The idea is not a far fetched as it used to be. This is not limited to active devices either. Passive devices, such as doors and windows can be connected as well. They will be talking to us and each other. All of this is the early stages of a programmable world and will take some time to sort out.

We know from Moore’s Law that electronic devices get cheaper all the time. Here is a direct link to the article in Wired that describes how will be connecting sensors to physical devices and integrating them into the programmable world.

Device sensors generating data

Sensors and objects linked together.

According to the author, there are three stages. The first involves getting devices on the network. The prospect of generating, monitoring this data and then triggering events as a result will lead to applications not within our purview currently. And, this it not just for residential use; factory automation can be taken to a new level. Bringing the analog world into the digital age. Having a system to collect and arrange the data is required. Think about when every device in your house is connected to your Wi-Fi network. The second stage will have those devices on the network sync with each other – output from one device triggers an action in another. The third and final stage involves using these devices as a system or single platform. To make this work, we will need repeatable and consistent patterns. The first generation will be crude and will not handle exceptions well. We will get smarter about that and iterate on stimulus and response triggers.

We have experienced discrete pieces of connected devices into disparate networks already. Wireless (aka, WiFi), Bluetooth, Radio Frequency Identification (RFID) and Near Field Communications (NFC) are routinely used in security badges, printers, cameras, smart phones and tablets. The next iteration will be connecting them together.

There is a great potential here with the number of devices to connect in the trillions. A big number. Each device generating data based on stimulus. Management will require programming –  lots of programming and networking to make this all work together.  Big companies are looking at this including Qualcomm, Cisco, GE and IBM as well as start ups are working on this as well. Changes will be seen in the home, factory and the office.

To revise an old phrase, this isn’t your Father’s network. Devices with embedded sensors will generate a lot of chatter. Who is going to listen? Where will that data be stored? What standards will be needed for command and control? We will get that sorted that out and then look for the next challenge.

Unlimited mobile data plans are changing June 3, 2010

Posted by Edwin Ritter in Trends.
Tags: , , ,
add a comment

As the cell phone market matures, the major cell phone carriers are looking for ways to keep/increase revenue streams. We are witnessing this evolution in the latest move by AT&T to migrate away from unlimited data usage. If the usage based pricing model is successful, expect the other major carriers to follow with their own plans. Similar to pricing changes in the airline industry, this move will be watched closely by competitors and analysts. The NY Times has an article that talks about AT&T plan to use tiered pricing.  It is not surprising that companies are looking to introduce usage based pricing plans. The carriers are seeing revenue streams plateau or even drop on the current pricing for phone plans. So, carriers see growing demand for data access and that can easily be converted to a revenue stream. Price points can be tricky though. How much is cloud access worth to you? What part of your disposable income are you willing to spend on getting email anywhere, anytime? Checking stock prices or sports scores?

This is a clear signal that the market is sorting out and defining what types of consumers we are. There is no one data plan for all users. Think of business travelers, sales reps., soccer moms and college students – each has unique reasons to access the web for information. I expect there will be a price plan for each based on their needs and budget.

I don’t take full advantage of the flat rate services I use. I know that and it’s OK. I favor the flat rate plan for service. It is predictable and a constant. It also is up to me to use the service as much, or, as little as I need it. My cable company provides a flat rate for video and I hope it stays that way. I know about  cable offers with premium packages, like the NFL network. But, that is a niche, not the mass consumer.  As this plan rolls out, people will become more aware of their usage for a short time. However, I don’t think usage rates will change much. In fact, the usage may increase as people find more ways to access the cloud to stay connected.

On the technical side, I understand the argument but do not agree that a few users can ‘hog’ the bandwidth on the network. Come on AT&T, you’ve been in the network business a long time. You have automated processes and tools to balance and manage the traffic with equipment in the network that is smart and programmable. Routers and switches can be tuned to give priority to specific types of data packets. Load balancing is not just for the data center anymore and has been used in network management for a long time. I don’t buy the argument anymore that a few users impact the response rate for everyone. Even if less than 2% of your customers are ‘heavy traffic users’, the network response can be managed smarter. I hope folks managing data centers aren’t watching this and think it can be used with their service. Wouldn’t that be a change?

I think it is an opportunity missed by the carriers. Network data usage is sporadic for most people; it is not a constant demand. There are peak times and I’m sure the trends on network demand are charted and known. Think business hours and prime time hours – there is a predictable level of usage and different parts of the network can be idle at times. The network is not routinely saturated at all points with traffic causing congestion and slow downs.  We cannot use all the bandwidth provided 24X7. It comes down to capacity management based on demand aligned with load balancing.

Have all the ways to reduce cost in the current system been exhausted? Has all the waste and redundancy in process been removed? Perhaps an opportunity to use Six Sigma and lean manufacturing tools here…give me a call and we can talk about it. 😉