Stacks and Brands
The problem is the stacks are of a poor quality. What is worse, they suffer in the areas which are difficult to identify and almost impossible to fix by a typical product company. So trusting the underlying stacks, the vendors continue building products and then ship them, exposing themselves to potentially significant brand damages.
If a company X ships a product based on a silicon / stack vendor A or B, and that product fails to perform, it is the X brand that suffers, as neither A nor B brands are identified. X very often has no competencies to judge the performance (or other hidden problems), as they are pushed by A's or B's sales people pitching the products datasheets. The datasheets have a list of features and nobody asks about the performance of the software stack, somehow assuming it all works.
One industry where this has become clear and unveiled is the mobile carrier industry. An early 3G/HSPA carrier told me some time ago "we allow only phones with Qualcomm chips on our network". This was independently confirmed later by an engineer at Ericsson mobile modem business unit, who told me they withdrew from the modem business, as they were always lagging behind Qualcomm in so called modem attach rate. The attach rate is responsible for dropped / interrupted calls. Mobile carriers knew that and to ensure a good experience on their networks, they tried to ban (or at least discourage) phones using chips with poor attach rates.
My first experience with a poor quality stack was back in 2014, when we selected a Bluetooth Low Energy SoC from a Tier-1 vendor and one of the key rationale for that choice was they were offering secure (encrypted) OTA (Over-The-Air) firmware update. Several months down the road we discovered that they had a design flaw leading to the security nonce being reused. So with a bit of effort an encrypted firmware sent over the air could be cracked, tampered with etc.
Fortunately we had the competences to fix that. At a significant cost. But again, many do not or even are not aware of such flaws.
Recently I have had an opportunity to put my hands down on several Bluetooth mesh devices running firmware based on standard stacks delivered by the chipset vendors. All three had significant flaws.
- One was simply resetting itself spontaneously in the presence of random radio traffic in the air. We notified the vendor and they acknowledged the issue but that did not stop them from launching the product.
- The second one had a very poor implementation of the relay feature. Relay is the fundamental feature in a mesh networking node: receive a message and retransmit it further. This relay was dropping 4 out of 5 messages, relaying only 20% of the traffic.
- The third one suffered in a similar area - the relaying seemed to be working fine until a phone connected to that device, at which point it started choking completely.
Now imagine somebody builds a network which includes devices suffering from the problems indicated above. What will be the result? All sorts of network reliability problems, disappointed customers and perhaps the contractor even going out of business.
One remedy here is to pick devices from leading brands, known for their quality. A big brand wants to protect the brand value. So it is unlikely they will be offering underperforming products. Apple may not be happy to pay Qualcomm the premium price for the mobile chips and software, but they would be even less happy if iPhones were suffering from dropped calls. IOW buying an Apple product is an implicit guarantee of performance. There are a hundreds of billions of dollars in company value at stake.
At Silvair, we take performance and quality very seriously. This is our brand value. It has been noticed and appreciated by all our ecosystem partners. In Bluetooth mesh networking Silvair means top quality, top performance and makes the other brands stronger, protected and secure. It is also the brand that supports Bluetooth mesh as a standard that people trust and rely on. As one of the co-founders I'm extremely happy to have quality as our primary company value. Proven.
Comments
Post a Comment