Concatenated SMS

Concatenated SMS, also known as concatenated short messages, is a feature of the Short Message Service (SMS) protocol that allows the transmission of text messages exceeding the standard single-message character limit by splitting the content into multiple segments, each transmitted as an individual SMS, and reassembling them in order at the receiving device using header information.^[1] This method addresses the inherent length restrictions of SMS, which are typically 160 characters for GSM-7 encoded messages or 70 characters for UCS-2 encoded messages, enabling longer communications such as detailed alerts, promotional content, or multi-part instructions without requiring alternative channels.^[2] Defined in the 3GPP technical specification TS 23.040, the process relies on a User Data Header (UDH) containing a reference number, sequence number, and total segment count to link and reconstruct the parts correctly.^[1] The segmentation reduces the payload per part to accommodate the UDH—typically 6 octets—resulting in 153 characters available per segment for GSM-7 encoding, 134 octets for 8-bit data, or 67 characters for UCS-2, with a theoretical maximum of 255 segments per message, though practical limits imposed by carriers and devices often cap it at 4 to 10 parts to ensure reliable delivery and reassembly.^[3]^[4] For instance, a full concatenated message in GSM-7 can reach up to 39,015 characters across all segments, but exceeding carrier thresholds may lead to fragmentation issues or increased costs, as each segment is typically billed separately by mobile networks.^[1]^[5] Support for concatenation is mandatory in modern GSM, UMTS, and LTE networks per 3GPP standards, ensuring compatibility across devices that adhere to the protocol, though older or non-compliant handsets may receive segments as independent messages.^[1] In practice, concatenated SMS is widely used in enterprise messaging applications, such as two-factor authentication codes, banking notifications, and marketing campaigns, where brevity alone suffices but extended detail enhances user experience; however, it requires careful encoding management to avoid unexpected splits due to special characters triggering UCS-2 mode and halving the per-segment capacity.^[2] The feature also integrates with SMS compression mechanisms outlined in the same 3GPP specification, allowing further optimization for longer payloads by applying headers and footers across segments.^[1] While effective for legacy mobile networks, concatenated SMS remains relevant even in 5G environments as a fallback for global interoperability, though it is increasingly supplemented by richer messaging protocols like RCS for multimedia support.^[5]

SMS Basics

Standard SMS Format

The Short Message Service (SMS) functions as a store-and-forward service in Global System for Mobile Communications (GSM) networks, where messages are temporarily stored at a Service Centre (SMSC) before being forwarded to the recipient, utilizing Signaling System No. 7 (SS7) for signaling between network elements such as Mobile Switching Centers (MSCs).^[6] SMS messages in Protocol Data Unit (PDU) mode, the standard binary format for transmission, consist of several core components defined in the Short Message Transfer Protocol (SM-TP). The SMSC address identifies the service center responsible for message handling, formatted as an E.164 number. Sender and recipient addresses are encoded in the TP-Originating Address (TP-OA) for incoming messages and TP-Destination Address (TP-DA) for outgoing ones, each spanning 2 to 12 octets and including type-of-number and numbering-plan indicators. The TP-Protocol Identifier (TP-PID), a single octet, specifies any higher-layer protocol or telematic service interworking, defaulting to 0x00 for basic short message transfer. The TP-Data Coding Scheme (TP-DCS), also one octet, indicates the message's data coding, alphabet, and language for proper interpretation. The TP-User Data Length (TP-UDL) denotes the size of the subsequent user data field in septets (for 7-bit encoding) or octets (for 8-bit or 16-bit), while the TP-User Data (TP-UD) carries the message payload itself.^[6] The default TP-UD payload is limited to 140 octets in a single SMS. When using the 7-bit GSM 03.38 default alphabet, this equates to 160 characters, achieved by packing 7 bits per character across 8-bit octets (e.g., 8 characters occupy exactly 7 octets without padding).^[6]^[7] This alphabet supports basic Latin characters, numbers, and common symbols suitable for Western European languages. For broader international support, UCS-2 encoding (per ISO/IEC 10646) is available, using 16 bits per character and thus fitting up to 70 characters in the 140-octet limit.^[7] In the standard format, the TP-UD directly contains the message text, though it may optionally prepend a User Data Header for certain extensions.^[6]

Character Limitations in Single SMS

A standard Short Message Service (SMS) message in Protocol Data Unit (PDU) mode is constrained to a fixed limit of 140 octets for the user data field within the TP-UD (Transport Protocol User Data).^[8] This limit arises from the underlying GSM/UMTS signaling structure, where the total SMS-DELIVER PDU can span up to 159 octets, allocating 140 octets specifically for the TP-UD payload.^[8] In cases where additional headers are included within the user data—such as for enhanced messaging features—the available payload space is reduced; for example, a User Data Header of 6 octets for concatenation reduces the effective payload to 134 octets.^[8] The actual number of characters that can fit within this 140-octet limit varies depending on the encoding scheme, as defined by the TP-DCS (Transport Protocol Data Coding Scheme) field in the SMS PDU.^[8] By default, SMS uses 7-bit packing with the GSM 7-bit default alphabet (as specified in 3GPP TS 23.038), allowing up to 160 characters per message, since 140 octets equate to 1,120 bits, and 1,120 / 7 = 160 characters exactly.^[8] For 8-bit binary data encoding—indicated by TP-DCS bits for uncompressed 8-bit user-defined coding—the limit drops to 140 octets of raw data, suitable for binary content but not character-based text.^[8] In 16-bit UCS-2 mode for Unicode support (e.g., for non-Latin scripts), each character requires two octets, restricting the message to 70 characters (140 / 2 = 70).^[8] The TP-DCS octet directly influences these encoding choices: bits 3-5 specify the coding group, with 000 indicating the default 7-bit GSM alphabet (packed for efficiency), 001 for unpacked 7-bit (treated as 8-bit, reducing capacity), 010 for 8-bit data, and 011 for UCS-2.^[8] This scheme ensures compatibility across networks but enforces trade-offs; for example, opting for unpacked 8-bit mode instead of packed 7-bit sacrifices the full 160-character capacity for broader character support without packing overhead.^[8] When a message mixes character sets—such as including emojis or non-GSM 7-bit characters (e.g., accented letters beyond the basic set or symbols like € in extended tables)—the entire message typically switches to UCS-2 encoding to maintain integrity, halving the capacity to 70 characters. For instance, a 100-character English message using only GSM 7-bit fits entirely, but adding a single emoji (encoded as two UCS-2 units) forces UCS-2 for the whole, potentially splitting it across segments if exceeding 70 characters. This reduction underscores the payload constraints that limit expressive content in single SMS messages.^[8]

Concept of Concatenated SMS

Definition and Purpose

Concatenated SMS is a technique that enables the transmission of longer text messages by dividing a single message into multiple segments, each sent as an independent standard SMS, which are then reassembled on the receiving device using a common reference identifier. This method allows mobile networks to handle content beyond the capacity of a single SMS without requiring alternative protocols.^[8] The primary purpose of concatenated SMS is to overcome the 160-character limitation of traditional single SMS messages, facilitating applications such as emergency alerts, promotional newsletters, or detailed multimedia descriptions that demand more space while avoiding the higher costs and complexity of MMS. It supports efficient delivery of essential business information, ensuring recipients view the full message as a cohesive unit rather than fragmented parts.^[2]^[5] This capability emerged in the late 1990s amid the evolution of Global System for Mobile Communications (GSM) standards, driven by increasing demand for enhanced mobile messaging features beyond basic short texts. Specified initially in GSM Phase 2+ enhancements, it was formalized in documents like ETSI GSM 03.40 (version 5.3.0, July 1996), providing a backward-compatible extension to early SMS deployments.^[6]^[8] A significant advantage of concatenated SMS lies in preserving the affordability and universal compatibility of conventional SMS infrastructure, while theoretically enabling messages up to 39,015 characters long in the 7-bit GSM default alphabet encoding (across 255 segments of 153 characters each, after accounting for segmentation overhead). This scalability addresses the restrictive 160-character boundary of single SMS, extending usability for diverse communication needs.^[8]

Message Length Extensions

Concatenated SMS enables the transmission of messages exceeding the 160-character limit of a single SMS by dividing the content into multiple segments, each carrying a portion of the payload while adhering to the 140-octet maximum for the TP-User Data (TP-UD) field in GSM/UMTS networks.^[8] The effective message length is extended through this segmentation, with the total capacity determined by the number of parts and the payload available per part after accounting for the User Data Header (UDH).^[8] The maximum number of parts in a concatenated message is 255, constrained by the 8-bit reference number in the UDH, which serves as a unique identifier for the segments and ranges from 0 to 255 (modulo 256).^[8] For the standard 8-bit reference concatenation (IEI=00), the UDH consists of 6 octets: 1 octet for the UDH length indicator (UDHL=05), followed by the 3-octet information element (IEI, length, and data including reference number, total parts, and sequence number).^[8] This leaves 134 octets for the payload in the 140-octet TP-UD field.^[8] Payload capacity varies by encoding scheme. In 7-bit default alphabet mode (GSM 7-bit), the 134 payload octets accommodate 153 characters, as each septet (7 bits) represents one character and the octets are packed accordingly (134 × 8 / 7 ≈ 153).^[8] In 8-bit data mode, the payload is exactly 134 octets.^[8] For UCS-2 (16-bit Unicode), the 134 octets support 67 characters, with each character requiring 2 octets.^[8] The total message capacity is calculated as the number of parts multiplied by the per-part character count; for example, 8 parts in 7-bit mode yield 8 × 153 = 1,224 characters.^[8] Several factors can reduce the effective length beyond the baseline UDH overhead. The inclusion of the 6 UDH octets inherently shortens the payload compared to single-SMS transmission.^[8] Variable encoding choices, such as switching to UCS-2 for non-Latin scripts, halve the character capacity per part relative to 7-bit mode.^[8] Additionally, optional elements in the UDH—such as application port addressing or timestamps—increase the UDH size (e.g., adding 3–5 octets per element), further reducing the payload to as few as 133 or fewer octets per part.^[8]

Technical Implementation

User Data Header (UDH)

The User Data Header (UDH) is a protocol element within the Short Message Service (SMS) that enables concatenation by embedding metadata at the beginning of the user data field in each message segment. It is indicated by setting the TP-User-Data-Header-Indicator (TP-UDHI) to 1 in the SMS Transfer Protocol Data Unit (TPDU), allowing the receiving device to recognize and process the header for reassembly. The UDH is optional and variable in length, but for basic concatenation, it consumes 6 octets per segment, reducing the available space for actual message content from the standard 140 octets to 134 octets in 7-bit default alphabet encoding.^[8] The UDH begins with a single octet specifying its length (UDHL), excluding the UDHL octet itself, followed by one or more Information Elements (IEs). Each IE consists of an Information Element Identifier (IEI) octet, an Information Element Data Length (IEDL) octet, and the corresponding data. For concatenated SMS, the relevant IEI for standard 8-bit reference numbering is 00, which identifies the concatenation data and is used when the message is split into up to 255 parts. This IE structure is: IEI (00), IEDL (03), an 8-bit reference number (0-255, unique across all segments of the message), an 8-bit total parts count (1-255), and an 8-bit sequence number (1 to total parts, indicating the segment's position). A less common IEI of 08 supports 16-bit reference numbering (0-65535) for scenarios requiring more unique identifiers, though total parts remain limited to 255; its structure includes IEI (08), IEDL (04), a 16-bit reference number, total parts, and sequence number.^[8] The UDH is prepended directly to the TP-User Data (TP-UD) field, ensuring that all segments share the same reference number while differing in sequence numbers to facilitate ordered reassembly. While the UDH can include other IEs, such as those for application port addressing (IEI 04 for 8-bit ports or 05 for 16-bit ports), the focus for concatenation is solely on the reference, total parts, and sequence elements to link segments without altering the core message flow. This design, as defined in 3GPP TS 23.040, ensures compatibility across GSM, UMTS, and later networks while minimizing overhead.^[8]

Segmentation and Reference Numbers

In concatenated SMS, a long message exceeding the single SMS payload limit is segmented into multiple shorter parts by the originating entity, such as a mobile station or service center, to fit within the 140-octet user data field of each SMS transfer protocol data unit (TPDU). The segmentation algorithm divides the original message payload into chunks sized according to the encoding scheme, accounting for the overhead introduced by the User Data Header (UDH) that contains concatenation metadata. For default GSM 7-bit encoding, each segment carries up to 153 characters; for 8-bit data encoding, up to 134 octets; and for UCS-2 (16-bit Unicode) encoding, up to 67 characters. This adjustment ensures that the UDH—typically 6 octets for an 8-bit reference number—does not exceed the available space, with the originating entity calculating the exact payload size per segment to avoid truncation.^[2] A unique reference number is assigned to all segments of a single concatenated message to enable identification and reassembly, generated by the sender using a random value or incrementing counter to minimize collision risks. This number is an 8-bit value (0-255) in the standard format, occupying 1 octet, or a 16-bit value (0-65535) in the extended format for higher uniqueness across high-volume traffic, occupying 2 octets. The reference number remains identical across all parts of the message and is included in the UDH of each segment, combined with the originating and service center addresses for complete disambiguation. Each segment is further distinguished by a sequence number indicating its position within the overall message, ranging from 1 for the first part to the total number of parts (up to 255, limited by the 8-bit field). The total number of parts is also specified in the UDH, allowing the receiver to track completeness. These numbers are encoded in 1 octet each within the UDH information element data. When the total message length does not divide evenly into the per-segment payload limits, the final segment is shorter than preceding ones, carrying only the remaining content without requiring padding to match the maximum size. This approach optimizes transmission efficiency while ensuring all segments except the last are uniformly sized for the chosen encoding.

Sending and Receiving Process

Encoding and Transmission

In the encoding process for concatenated SMS, each message segment incorporates a User Data Header (UDH) within the TP-User Data (TP-UD) field to facilitate linkage, with the TP-User Data Header Indicator (TP-UDHI) bit set to 1 to signal its presence.^[9] The UDH typically uses an 8-bit reference format (IEI=00), consisting of a 1-octet UDHL followed by the reference number, total number of segments, and sequence number for the current segment, allowing up to 255 parts; a 16-bit reference (IEI=08) extends this to 65,535 segments for larger messages.^[9] The TP-User Data Length (TP-UDL) is set to the total length of the TP-UD in octets (for 8-bit data) or septets (for 7-bit default alphabet), explicitly including the UDH overhead, which reduces the effective payload to 134 octets or 153 septets per segment, respectively.^[9] The sender device, whether a mobile handset or an application interfacing with a GSM modem, generates these encoded segments from the original segmented message.^[10] In modem-based systems, this is achieved using AT commands defined in GSM 07.05, particularly the +CMGS command in PDU mode, where the application supplies the full hex-encoded TPDU—including the UDH—for each segment, followed by a termination character like .^[10] For example, +CMGS= prompts entry of the PDU string, enabling precise control over UDH placement and segment-specific details like the reference number and sequence.^[10] Once encoded, the segments are transmitted as independent SMS-SUBMIT PDUs through the Short Message Service Centre (SMSC), with each part routed separately and the TP-Message Reference (TP-MR) incremented to distinguish them, though arrival order is not guaranteed due to network variability.^[9] The network path begins at the Mobile Station (MS), which forwards the PDUs to the Mobile Switching Centre (MSC) or Serving GPRS Support Node (SGSN) for initial handling.^[9] From there, the MSC/SGSN relays them to the SMSC via the Mobile Application Part (MAP) protocol over SS7 signaling, using operations like ForwardShortMessage for store-and-forward delivery.^[9] The SMSC then dispatches each segment toward the destination without inherent reassembly, preserving their autonomy in transit.^[9]

Reassembly on Receiver Side

Upon receipt of an SMS message, the receiving device first parses the User Data Header (UDH) within the Transport Protocol Data Unit (TPDU) to detect concatenation. This is achieved by examining the first Information Element Identifier (IEI) in the UDH, which is set to '00' to indicate that the message is part of a concatenated short message.^[8] The receiver then matches incoming segments to the same original message using a reference number embedded in the UDH of each part. This reference number, an 8-bit value (0-255) operating modulo 256, remains constant across all segments and is used in combination with the originating address and protocol identifier (TP-PID) to group segments belonging to the same message. Segments are grouped accordingly, while duplicates are processed individually to avoid redundant handling, and orphaned parts—those without matching counterparts—are discarded if they cannot be completed.^[8] To ensure correct ordering, the receiver sorts the matched segments based on their sequence numbers, which are 8-bit values ranging from 1 to 255 and increment sequentially for each part. The UDH also includes a maximum number field (0-255) indicating the total segments expected, allowing the receiver to verify completeness; invalid sequence numbers (e.g., 0 or exceeding the maximum) result in segment discard.^[8] Reassembly involves buffering any out-of-order segments—arising from independent transmission paths—and concatenating the user data payloads after stripping the UDH from each, starting from the appropriate septet or octet boundary to form the original message. Buffering persists until all segments arrive or the message is deemed incomplete, with reassembly logic relying on the reference and sequence details for accurate reconstruction.^[8] On mobile devices, this process is typically managed transparently by the phone's firmware or the underlying SMS stack, without user intervention. For example, in Android, the telephony framework processes incoming PDUs via the SmsMessage class, which exposes methods like createFromPdu to access segment data, while the system handles grouping and reassembly before delivering the full message to applications.^[11]

Standards and Compatibility

3GPP and ETSI Specifications

The formal standardization of concatenated SMS is primarily defined in the 3GPP Technical Specification TS 23.040, titled "Technical realization of the Short Message Service (SMS)," which has governed the protocol since its initial release in 1999. This specification details the use of the User Data Header (UDH) for concatenation in section 9.2.3.24, enabling the segmentation and reassembly of longer messages across multiple SMS units.^[8] Prior to 3GPP's formation, the European Telecommunications Standards Institute (ETSI) introduced concatenated SMS as part of the Global System for Mobile Communications (GSM) phase 2+ enhancements in GSM 03.40, version 5.3.0, released in July 1996.^[6] This ETSI specification laid the groundwork for point-to-point SMS operations, including the initial 8-bit reference number mechanism in the UDH for linking message segments.^[6] Support for 16-bit reference numbers for concatenated SMS was introduced in 3GPP TS 23.040 version 2.0.0, released in June 1999, as defined in section 9.2.3.24.8 with Information Element Identifier (IEI) 08, allowing for up to 65,536 unique references to accommodate larger message volumes.^[12] Subsequent releases have aligned concatenated SMS with IP Multimedia Subsystem (IMS) enhancements, such as those in TS 24.341 for SMS over IP, to support multimedia messaging in 4G and 5G networks. Key protocol elements include the TP-Message Type Indicator (TP-MTI) in section 9.2.3.1 of TS 23.040, where the binary value 01 denotes SMS-SUBMIT (from mobile station to service center) and 00 denotes SMS-DELIVER (from service center to mobile station), both requiring the TP-UDHI bit to be set for UDH inclusion in concatenated messages.^[8] Conformance to UDH requirements ensures interoperability, mandating that receiving entities reassemble segments based on the reference number, total segments, and sequence position as per section 9.2.3.24.1 (8-bit) or 9.2.3.24.8 (16-bit).^[8]

Support in Different Networks

Concatenated SMS enjoys full support in GSM, UMTS, and LTE networks, with compatibility established since the introduction of 2G (GSM) systems. The 3GPP specifications mandate that core network elements, such as the Mobile Switching Center (MSC) and Visitor Location Register (VLR) in GSM and UMTS, as well as the Mobility Management Entity (MME) in LTE (EPS), handle segmented messages without alteration, ensuring segments are routed appropriately to the Short Message Service Center (SMSC) for forwarding. This support extends to 5GS in later evolutions, maintaining backward compatibility through standardized protocols like MAP in earlier generations and Diameter in LTE. In 5G networks (5GS), concatenation is supported via SMS over NAS as defined in 3GPP Release 15 and subsequent versions, ensuring compatibility with legacy SMS protocols.^[13] In CDMA networks, support for concatenated SMS is partial and relies on adaptations to the IS-41 signaling protocol, as defined in 3GPP2 standards. The core SMS specification (C.S0015) limits messages to single parts, with fields like MSG_NUMBER and NUM_MSGS required to be 1, preventing native multi-part handling; however, some carriers implement proprietary extensions or workarounds at the SMSC level to enable segmentation for intra-network delivery.^[14] Interworking with GSM-based networks often requires additional adaptations, but full concatenation may not be preserved across boundaries.^[15] Device compatibility for concatenated SMS is robust in modern smartphones, where reassembly of segments using the User Data Header (UDH) is a required feature. Modern iOS and Android devices support concatenation, with iOS handling via the Messages app and Android through the SmsMessage class in its telephony framework. In contrast, older feature phones often lack full UDH processing, potentially dropping segments or displaying them as separate, garbled messages with cryptic characters from the header.^[16] Interoperability challenges arise during roaming between networks, particularly if the visited network's SMSC strips or fails to preserve the UDH, leading to incomplete reassembly at the recipient device. This issue is common in cross-technology roaming (e.g., GSM to CDMA), where protocol mismatches in SS7-based handoffs can disrupt segment linking, though guidelines recommend SMSCs maintain UDH integrity for end-to-end delivery.^[16]

Limitations and Best Practices

Delivery Reliability Issues

One significant challenge in concatenated SMS delivery is the independent transmission of each segment, which lacks any network-level guarantee of sequential order or complete arrival. As defined in the 3GPP technical specification for SMS, segments are routed separately through service centers, potentially via different paths, leading to out-of-order arrival at the receiving device. This can delay reassembly, as the mobile station must buffer incoming parts until the full set is received, sometimes extending wait times to several minutes depending on network conditions.^[9] Part loss further compromises reliability, as individual segments can fail to deliver due to factors such as network congestion, resource shortages, or maximum retry limits being exceeded, without triggering a retry for the entire message. Analysis of nationwide SMS traces reveals an overall delivery failure rate of approximately 5.1% under normal conditions, primarily from retry exhaustion (3.5%) and message expiration (1.6%), with congestion events like peak traffic periods exacerbating losses through increased paging failures and channel contention. Since segments are treated as standalone messages by the network, the loss of even one part results in an incomplete concatenated message, as no built-in mechanism exists for end-to-end acknowledgment or holistic retransmission.^[9]^[17] In practice, single-segment SMS delivery rates typically range from 95% to 98%, but the multiplicative nature of multi-part success—requiring all segments to arrive—reduces overall reliability for longer messages, highlighting the compounded risk in concatenated scenarios. When reassembly fails, fallback behaviors vary by device: some display partial content from received segments, while others show an error notification, as there is no standardized end-to-end confirmation for the full message.^[18]^[19]

Optimization Tips

To enhance the performance and reliability of concatenated SMS, limit the message to eight or fewer segments, as this aligns with the maximum supported by most carriers and minimizes the risk of partial losses during transmission.^[20] Prefer 7-bit GSM-7 encoding over UCS-2 when possible, enabling up to 153 characters per segment for standard Latin text and thereby maximizing overall payload efficiency without invoking less optimal Unicode handling.^[20]^[21] Effective reference management in the User Data Header (UDH) requires assigning unique, sequential reference numbers to each concatenated message across sending sessions, which prevents collisions that might cause segments from different messages to be incorrectly reassembled on the receiving device.^[4] For validation, test concatenated SMS using AT commands on cellular modems or via SMS gateway APIs to confirm proper segmentation and reassembly; monitoring Short Message Service Center (SMSC) logs provides insights into potential failures, such as undelivered segments.^[22]^[23] When messages surpass 1600 characters or incorporate rich media like images, opt for MMS as an alternative, which eliminates strict character limits and supports multimedia; for application-specific needs, binary SMS offers greater efficiency by transmitting non-text data in 8-bit format, reducing overhead compared to text-based encoding.^[24]^[25]